Selection of Connective Tissue Progenitors Based

SELECTION OF CONNECTIVE TISSUE PROGENITORS BASED

ON CELL-ASSOCIATED HYALURONAN FOR ENHANCED

BONE REGENERATION

By Tonya Caralla

Submitted in partial fulfillment of the requirements for the

degree of Doctor of Philosophy

Department of Biomedical Engineering

CASE WESTERN RESERVE UNIVERSITY

August 2012 Case Western Reserve University

School of Graduate Studies

We hereby approve the dissertation of

Tonya Caralla

Candidate for the Doctor of Philosophy degree*.

Committee Members:

Steven Eppell, chair and academic advisor

George Muschler, research advisor

Horst von Recum

Vincent Hascall

Maciej Zborowski

Defense date: November 4th, 2011

*We also certify that written approval has been obtained for any proprietary

material contained therein.

1 Dedication:

There have been many who have contributed to my growth; both as a researcher and a person. I would like to take a moment to acknowledge those who have shaped me into the person I am today – this dissertation is for you too.

To God, for bestowing on my countless blessings, a loving family, supportive friends, an amazing husband, and a giant brain.

To my father Louis Caralla, who has always pushed me to be the best and brightest version of myself, who pushed me toward engineering and graduate school, and who left this earth last year before he could see this doctorate completed. I held his hand as he passed away, and I hope he will be holding mine as I go through this dissertation defense.

I love you Dad, and I miss you every day.

To my mother Becky, who has always been there to nurture me and listen to me whine before telling me to suck it up. It‟s a fine line, and she‟s walked it well. Thank you for always telling me that I could do anything that I wanted, for tackling the unglamorous job of raising me (especially as a surly know-it-all teen), and for bestowing on me both

Barbies and Legos.

To my sister Alyssa, who was my creative outlet and fellow troublemaker as we grew up moving around the Midwest. Not just anyone would play Legos with me for hours and then cheerfully dig up the neighbor‟s yard together. I admire your brain for being the opposite of mine, your patience, and your self-confidence at literally everything you do. I still remember that you ruined my sticker collection though.

2 To my loving husband Joe, for being my rock through hard times and a constant source of joy and backrubs. I would not have gotten through the last 4 years without your love and support, and I am so excited to spend the rest of my life with you.

To my friends, who have become my family in Cleves and who offer excellent advice and unconditional love. They are patient, fantastic listeners, and always there when I need them (whether or not they are actually there in person). Leslie, Jones, Sam, Jen, Brandy,

Maria, Amanda, Aniq, Kanger – love you guys!

To my co-workers at the CCF, who are a font of knowledge and are always willing to take time to help me think through experiments, especially Cynthia, Chris, Vivek,

Viviane and Sandra.

To my committee, especially my advisor Dr. Muschler, who have pushed me to think deeper and become a better scientist, engineer, and researcher.

3 Table of Contents:

Committee Approval Sheet……………………………………………………………..1 Dedication……….………………………………………………………………………2 Contents ...... 4 List of Tables ...... 10 List of Figures ...... 11 Abstract………………………………………………………………………………….13 Executive Summary…………………………………………………………………….15

Chapter 1 Introduction to the Bone Tissue Engineering Paradigm, Connective Tissue Progenitor Cells, and Hyaluronan...... 21 1.1. Clinical need for bone grafts……………………………………………………...21 1.2. Bone tissue engineering paradigm………………………………………………..21 1.3. Materials used for bone grafts…………………………………………………....22 1.4. Need for osteogenic cells……………………………………………..……….…24 1.5 Osteogenic cells present in bone marrow………………………………………....26 1.6 Review of the properties of hyaluronan…………………………………………..28 1.6.1 Composition………………………………………………………………....…28 1.6.2. Size and physical properties………………………………………………....…29 1.6.3. Degradation………………………………………………………………….….30 1.7 HA configuration in vivo……………………………………………………..…...31 1.8 Hyaluronan in disease states: inflammation, wound healing, and angiogenesis …35 1.9 Hyaluronan in the bone marrow…………………………………………………..36 1.10 Concept of a stem cell niche……………………………………………………....37 1.11 Hematopoietic stem cell niche……………………………………………...……..37 1.12 Evidence for CTP niche……………………………………………………….…..39 1.13 References………………………………………………………………………....43

Chapter 2 Introduction to Cell Separation Methods and Cell Surface Marker Characterization of CTPs……………………………………...………………………...64 2.1 Use of autologous bone marrow in bone defects………………………………..…64 2.2 Density gradient separation methods for cell separation…………………………..65 2.3 Selective Retention methods for cell separation…………………………………...66 2.4 Fluorescence activated cell sorting (FACS) for cell separation ………………...…68 2.5 Magnetic separation technologies for cell separation……………………………...69 2.6 Magnetic Theory………………………………………………………………..….69 2.7 Types of Magnetism……………………………………………………………….71 2.8 Clinical Application of Magnetic Separation Systems………………………….…73 2.9 Advantages of commercially available magnetic separation systems……………..74 2.10 Geometry of the Magnet………………………………………………………...…76 2.11 Cell surface markers used to select CTPs……………………………………….…77 2.11.1 Selection based on STRO-1……………………………………………………81 2.11.1 Selection based on CD105…………………………………………………..…82 2.11.3 Separation based on CD49a……………………………………………………82 2.11.4 Separation based on CD73……………………………………………………..83

4 2.11.5 Selection based on CD271……………………………………………………..83 2.11.6 Separation based on CD146……………………………………………………84 2.12 References………………………………………………………………………….86

Chapter 3 Hyaluronan as a novel marker for rapid selection of connective tissue progenitors……………………………..…………………………………………...……….93 3.1 Abstract………………….…………………………………………………………93 3.2 Introduction……………………………………………………………………..…94 3.3 Methods……………………………………………………………………………96 3.3.1 Cell Sources……………………………………………………………………...96 3.3.2 Magnetic Separation……………………………………………………………..97 3.3.3 Hematopoietic Progenitor Cells (HPC) Assay…………………………………..98 3.3.4 Colony Forming Unit Assay for CTP-Os………………………………………..98 3.3.5 Image Analysis………………………………………………………………..…99 3.3.6 Flow Cytometry………………………………………………………………...100 3.3.7 Statistical Analysis……………………………………………………………...102 3.4 Results…………………………………………………………………………….102 3.4.1 Cell and CTP Yield……………………………………………………………..102 3.4.2 Purity and Recovery…………………………………………………………….104 3.4.3 Four color flow cytometric analysis……………………………………………106 3.4.4 Selection of Colony Forming HPCs……………………………………………107 3.4.5 Selection of Colony Forming CTPs……………………………………………109 3.4.6 Biological Performance of Colonies Derived form CTPs Isolated in the HA+++, HA+, and HA-…………………………………………………………………………...109 3.4.6.1 Proliferation…………………………………………………………….109 3.4.6.2 Alkaline Phosphatase Activity………………………………………….110 3.4.6.3 Retained versus Newly Synthesized Hyaluronan………………………112 3.5 Discussion ………………………………………………………………………..113 3.6 Acknowledgments………………………………………………………………..116 3.7 References………………………………………………………………………..116

Chapter 4: A Model of Retention of Magnetized Cells in the EasySep Magnetic Separation System………………………………………………………………….……………119 Abstract………………………………………………………………………………... 119 4.1 Introduction……………………………………………………………………….120 4.2 Materials and Methods…………………………………………..………………. 122 4.2.1 Magnetic Labeling in the EasySepTM System…………………….…………… 122 4.2.2 Geometry of the EasySep Magnet……………………….……………………. 123 4.2.3 Magnetic Flux Density of the EasySep Magnet….…….………..……………..124 4.2.4 Testing with standardized Micromod beads……………………..……………. 125 4.2.5 Generation of HA+++ cells for CTV analysis………………………..…………. 126 4.2.6 Cell Tracking Velocimetry………………….………………………………… 127 4.3 Theory and Calculations………………………………………………………... 128 4.3.1 Derivation of model…………………………………………………………... 128 4.3.2 Relating model equation to magnetophoretic mobility……………………….. 132 4.4 Results…………………………………………………………………………... 135

5 4.4.1 Standardized Micromod Beads……………………………………………….. 135 4.4.2 Cell Tracking Velocimetry (CTV)……………………………………………. 138 4.5 Discussion………………………………………………………………………. 142 4.6 Appendi…………………………………………………………………………. 144 4.6.1 Appendix A: Justification: Neglecting the Inertial Term…………………….. 144 4.6.2 Appendix B: Justification: Sedimentation << Magnetic Displacement……… 145 4.7 References………………………………………………………………………. 150

Chapter 5 Development of protocol for magnetic separation based on hyaluronan for enrichment of connective tissue progenitors……………………………………………152 5. Chapter Introduction…………………….…………………………………………152 5.1 Evaluation of a method for RBC removal: LymphoprepTM Separation………..…154 5.1.1 Introduction……………………………………………………………..………154 5.1.2 Methods…………………………………………………………………………156 5.1.2.1 Cell Source……………………………………………………………...156 5.1.2.2 Description of LymphoprepTM Protocols……………………………….156 5.1.2.3 Cell Culture Conditions………………………………………………...158 5.1.2.4 ColonyzeTM Image Acquisition and Analysis…………………………. 159 5.1.2.5 Statistical Analysis…………………………………………………….. 159 5.1.3 Results…………………………………………………………………………. 159 5.1.3.1 Cell Counts and Yield after LymphoprepTM Processing………………. 159 5.1.3.2 CTP Prevalence after LymphoprepTM Processing…………………….. 161 5.1.3.3 Total number of CTPs in each layer after LymphoprepTM Processing...162 5.1.4 Conclusion…………………………………………………………………….. 163 5.1.5 Discussion……………………………………………………………………... 164 5.2 Evaluation of a simplified single pass magnetic protocol against the previously established three pass protocol to minimize residence time in the magnet…………… 166 5.2.1 Introduction…………………………………………………………………… 166 5.2.2 Methods………………………………………………………………………... 166 5.2.2.1 Cell Source…………………………………………………………….. 166 5.2.2.2 Magnetic Separation Protocol using the EasySepTM system …………. 167 5.2.2.3 Hematopoietic Progenitor Cell (HPC) Assay…………………………. 168 5.2.2.4 Colony Forming Unit Assay for CTPs………………………………... 168 5.2.2.5 ColonyzeTM Image Acquisition and Analysis ………………………….169 5.2.2.6 Flow Cytometry……………………………………………………….. 169 5.2.2.7 Statistical Analysis…………………………………………………….. 170 5.2.3 Results……..……………………………………………………………………170 5.2.3.1 Cell and CTP yield and partitioning after magnetic separation………. 170 5.2.3.2 CTP Partitioning after magnetic separation…………………………… 171 5.2.3.3 Purity and Recovery assessed by flow cytometry…………………….. 173 5.2.3.4 Partitioning of Colony Forming HPCs ………………………………...175 5.2.3.5 Partitioning of CFU-M and CFU-E to MS-processed fractions………. 177 5.2.3.6 Prevalence of Colony Forming CTPs…………………………………. 178 5.2.3.7 Biological Performance of Colonies Derived form CTPs .....………… 179 5.2.3.7.1 Proliferation…………………………………………………… 179 5.2.3.7.2 Alkaline Phosphatase Staining…………………………………180

6 5.2.4 Discussion…………………………………………………………………….. 181 5.3 Effect of primary and secondary label concentration on the retention of magnetized cells and CTP prevalence in the HA-positive fractions……………………………….. 184 5.3.1 Introduction……………………………………………………………………. 184 5.3.2 Methods………………………………………………………………………... 185 5.3.2.1 Cell Source…………………………………………………………….. 185 5.3.2.2 Magnetic separation and variation of amount of labeling reagents…… 185 5.3.2.3 Cell Culture Conditions……………………………………………….. 186 5.3.2.4 ColonyzeTM Image Acquisition and Analysis…………………………. 187 5.3.3 Results…………………………………………………………………………. 187 5.3.3.1 Partitioning of Cells after MS processing……………………………... 187 5.3.3.2 Magnetophoretic moility of cells after MS processing………………... 188 5.3.3.3 CTP Prevalence………………………………………………………... 189 5.3.3.4 CTP Partitioning………………………………………………………. 190 5.3.4 Discussion……………………………………………………………………... 191 5.4 Effect of preincubation of labeling reagents to minimize the number of labeling steps and incubation time required to magnetically label the cells……………………. 194 5.4.1 Introduction…………………………………………………………………… 194 5.4.2 Methods………………………………………………………………………... 195 5.4.2.1 Cell Source…………………………………………………………….. 195 5.4.2.2 Magnetic Separation the EasySepTM system…………………………... 196 5.4.2.3 Cell Culture Conditions……………………………………………….. 197 5.4.2.4 Image Acquisition and Analysis………………………………………. 197 5.4.3 Results…………………………………………………………………………. 198 5.4.3.1 Cell Partitioning after MS processing…………………………………. 198 5.4.3.2 CTP Prevalence………………………………………………………... 198 5.4.3.3 Cells per colony……………………………………………………….. 199 5.4.3.4 Alkaline phosphatase expression……………………………………… 200 5.4.3.5 Partitioning of CTPs after magnetic separation……………………….. 200 5.4.4 Conclusion…………………………………………………………………….. 201 5.4.5 References………………………………………………………………………201

Chapter 6: Defining a Method for Loading of Magnetically Separated Cells onto Osteoconductive Scaffolds ……………………………………………………………..203 6. Chapter Introduction……………………………………………………………... 203 6.1 Evaluation of CTP rentention efficiency and selection ratio on the loading of HA+++ cells using selective retention (SR) on PLCL-TCP scaffolds…………………. 204 6.1.1 Methods……………………………………………………………………….. 205 6.1.1.1 Preparation of PLCL-TCP Scaffolds………………………………….. 205 6.1.1.2 Marrow aspiration and preprocessing…………………………………. 206 6.1.1.3 Labeling and Magnetic Separation using the EasySepTM system……... 206 6.1.1.4 Cell Culture Conditions……………………………………………….. 207 6.1.1.5 Selective retention loading of cells onto PLCL-TCP scaffolds……….. 207 6.1.2 Results………………………………………………………………………….208 6.1.2.1 Cell and CTP partitioning after EasySepTM magnetic separation…….. 208 6.1.2.2 Retention Efficiency after PLCL-TCP Scaffold Loading…………….. 209

7 6.1.2.3 Selection Ratio after PLCL-TCP Scaffold Loading…………………... 211 6.1.3 Conclusion…………………………………………………………………….. 212 6.2 Evaluation of CTP rentention efficiency and selection ratio on the loading of MS- processed cells using manual selective retention (MSR) or drip soaking (DS) on canine allograft matrix………………………………………………………………………… 213 6.2.1 Methods………………………………………………………………………... 213 6.2.1.1 Preparation of Mineralized Cancellous Allograft……………………... 213 6.2.1.2 Plasma and marrow collection and preprocessing…………………….. 214 6.2.1.3 Labeling and Magnetic Separation using the HMS…………………… 214 6.2.1.4 Loading of cells onto allograft scaffolds……………………………… 215 6.2.1.5 Cell Culture Conditions…………………………………………….…. 217 6.2.2. Results……………………………………………………………………….….218 6.2.2.1 Cell and CTP Partitioning after HMS Magnetic Separation…………... 218 6.2.2.2 Retention Efficiency for Cell and CTPs after DS and MSR loading….. 219 6.2.2.3 Selection Ratio after DS and MSR loading…………………………… 220 6.2.3 Conclusions……………………………………………………………………. 221 6.3 Chapter Discussion……………………………………………………………… 222 6.4 References……………………………………………………………………….. 228

Chapter 7: In vivo transplantation of autogenous marrow-derived cells following rapid intraoperative magnetic separation based on HA to augment bone regeneration.….. 231 7.1 Abstract………………………………………………………….……………… 231 7.2 Introduction…………………………………………………….……………….. 232 7.3 Methods……………………………………………………….………………… 234 7.3.1 Preparation of Allograft…………………………………….…………………. 234 7.3.2 Collection of heparinized Bone Marrow and Plasma …….……………………235 7.3.3 Magnetic Separation and Preparation of MCA Implants….………………….. 235 7.3.4 Canine Femoral Multidefect (CFMD) Model…………….…………………... 237 7.3.5 Assay for CTPs………………………………………………………………... 238 7.3.6 MicroCT Acquisition and Analysis…………………………………………… 239 7.3.7 Histological Analysis…………………………………………………………. 241 7.3.8 Statistical Analysis……………………………………………………………. 242 7.4 Results………………………………………………………………………….. 242 7.4.1 Comparison of hBMA and BCM…………………………………………….. 242 7.4.2 Comparison of the Magnetic Separation (MS) Fractions…………………….. 243 7.4.3 Comparison of Colony Metrics………………………………………………. 244 7.4.4 Accounting for the CTPs after MS processing……………………………….. 245 7.4.5 Allograft Loading…………………………………………………………….. 245 7.4.6 Outcome of In Vivo Implantation in the CFMD Model………….…………… 246 7.4.7 Histological Analysis…………………………………………………………. 248 7.5 Discussion………………………………………………………………………. 251 7.6 Acknowledgements……………………………………………………………... 258 7.7 References………………………………………………………………………. 259

Chapter 8: Conclusions and Future Work…………………………………………….. 263 8.1 Conclusions……………………………………………………………………... 263

8 8.1.1 HA as a novel marker for CTPs………………………………………………. 263 8.1.2 HA as a potential niche component for CTPs in vivo………………………… 263 8.2 Future work……………………………………………………………………… 264 8.3 Conclusions……………………………………………………………………... 267 8.3.1. The application of HA-positive CTPs for improved bone regeneration……… 267 8.4 Future work……………………………………………………………………… 268 8.4.1 Implementation of protocol changes………………………………………….. 268 8.4.2 The Chronic Segmental Defect (CSD) Caprine model……………………….. 269 8.4.3 Development of alternative methods for selection of CTPs based on HA……… 271 8.5 References………………………………………………………………………. 271

9 List of Tables:

Table 2.1 Cell surface markers used for the prospective isolation of native bone marrow- derived progenitor cells………………………………………………………………....80 Table 3.1 Analysis of 4 color flow cytometry on each cell fraction………………….. 106 Table 4.1: Parameters used for the calculation of the velocity of the magnetized cells……………………………………………………………………………………. 134 Table 4.2: Calculation for retention of standardized beads in the EasySepTM magnet: theoretical vs actual values……………………………………………………………. 137 Table A.1: Order of magnitude analysis……………………………………………… 145 Table B.1 Calculations for sedimentation displacement……………………………… 147 Table B.2 Calculations for magnetic displacement…………………………………… 148 Table B.3 Calculations for random thermal displacement……………………………. 149 Table 5.1 Systematic variation of staining parameters……………………………….. 186

10 List of Figures:

Figure 2.1 Example of staining sequence with EasySep reagents using HA...... …….…76 Figure 3.1: The ColonyzeTM software system………………………………….………100 Figure 3.2: CTPs are enriched in the HA+++ fraction…………………………….…….103 Figure 3.3: Cell and CTP partitioning after magnetic separation………………….…..104 Figure 3.4: Flow cytometric analysis of each fraction after magnetic separation…..…105 Figure 3.5: Prevalence of myeloid and erythroid progenitors…………………………108 Figure 3.6: Quantification of CTP prevalence………………………………………....109 Figure 3.7: Quantification of proliferation and differentiation………………………...111 Figure 4.1: Example of the staining sequence with EasySepTM reagents……………...123 Figure 4.2: View of the EasySepTM magnet and magnetic field lines………………... 124 Figure 4.3: Measurement of the magnetic flux density as a function of radial position………………………………………………………………………………….125 Figure 4.4: Force Diagram on the magnetized cell…………………………………… 128 Figure 4.5: Histogram of bone marrow aspirate obtained by Vi-Cell counter analysis..131 Figure 4.6: Percent of beads captured in magnet during varying residence times…… 138 Figure 4.7: Magnetophoretic mobility of the magnetized cells as measured by the CTV system…………………………………………………………………………………. 139 Figure 4.8: Images of the magnetically induced velocity and sedimentation………… 140 Figure 4.9: Magnetization and velocity of the magnetized cell as a function of radial position.………………………………………………………………………………... 141 Figure 4.10: Predicted time for the magnetized cells to be retained in the magnet…....142 Figure 5.1: Illustration of buffy coat LymphoprepTM density gradient separation…… 155 Figure 5.2: Fractions harvested after LymphoprepTM separation…………………….. 158 Figure 5.3: Partitioning of nucleated cells after LymphoprepTM preprocessing……… 161 Figure 5.4: CTP prevalence in each fraction after LymphoprepTM separation……….. 162 Figure 5.5: CTP partitioning after LymphoprepTM separation……………………….. 163 Figure 5.6: ColonyzeTM images illustrating the partitioning of CTPs ………………...172 Figure 5.7: Gylcophorin A negative cells………………...……………………………174 Figure 5.8: The non-erythroid cells are examined for hyaluronan expression……….. 175 Figure 5.9: Prevalence of myeloid progenitors (CFU-M)……………………………..176 Figure 5.10: Prevalence of erythroid progenitors (CFU-E)……………………………177 Figure 5.11: CTP prevalence in each fraction after MS processing………………….. 179 Figure 5.12: The number of cells per colony…………………………………………. 180 Figure 5.13: Alkaline phosphatase expression……………………………………….. 181 Figure 5.14: Partitioning of cells to the HA+ fraction ………………………………...188 Figure 5.15: Mean magnetophoretic mobility of the positively labeled cells………... 189 Figure 5.16: CTP Prevalence as a function of decreasing label concentration………. 190 Figure 5.17: Partitioning of CTPs after magnetic separation in the HA+ fraction…… 191 Figure 5.18: Cell partitioning after magnetic separation processing…………………. 198 Figure 5.19: CTP prevalence after magnetic separation standardized to the BCM…... 199 Figure 5.20: Cells per colony, standardized to the BCM……………………………... 200 Figure 6.1: PLCL-TCP “waffle” design scaffolds……………………………………. 205 Figure 6.2: Retention efficiency RE using BMA and HA+++ cells…………………… 211 Figure 6.3: Image of canine allograft chips used for loading of HA+++ cells………… 216

11 Figure 6.4: Illustration of manual selective retention and drip soaking methods……. .217 Figure 6.5: Cell partitioning after HMS separation………………………………….. .218 Figure 6.6: CTP Prevalence after HMS processing………………………………….. .219 Figure 6.7: Retention efficiency using DS or MSR methods………………………… 220 Figure 6.8: Allograft chips loaded by either DS or MSR methods…………………… 221 Figure 7.1: Micro-CT processing Technique…………………………………………. 240 Figure 7.2: CTP Prevalence after MS processing…………………………………….. 244 Figure 7.3: Results of MicroCT………………………………………………………. 247 + Figure 7.4: Histological staining of site grafted with MS/HA W2 cells………………. 248 Figure 7.5: Histological staining of site grafted with hBMA………………………… 249 Figure 7.6: Histomorphological analysis of soft tissue composition …………………250 Figure 7.7: Histomorphometric analysis of bone …………………………………….251 Figure 8.1: Co-expression of CD146 and HA on a colony after 6 days of culture……265

12 Selection of Connective Tissue Progenitors Based on Cell-associated

Hyaluronan for Enhanced Bone Regeneration

Abstract

TONYA CARALLA

The frequency of CTPs in bone marrow is quite low and this rarity highlights the need for clinically relevant selection strategies to optimize the impact of these cells in a graft site. This project represents the first exploration of the potential utility of using the extracellular matrix niche around a cell as a potential marker for positive or negative selection.

We hypothesized that hyaluronan represents a distinguishing feature of the in vivo niche of one or more subsets of the heterogeneous population of CTPs in bone marrow, and would therefore provide a means for enrichment of CTPs from a fresh bone marrow aspirate. HA-based magnetic separation resulted in the isolation of a cell fraction enriched in highly proliferative CTPs.

We also hypothesized that positive selection of CTPs based on hyaluronan (HA) will increase the efficacy of a bone graft in vivo due to increased concentration and prevalence of CTPs in a graft site. In order to test this hypothesis in vivo in a biologically relevant large animal model, we designed a new magnetic separation system to process

13 larger volumes of marrow and defined a preferred method (drip soaking) for loading of

HA-positive cells onto scaffolds.

The canine femoral defect model provides a method for comparison of bone grafting materials using four cylindrical bone defects in the canine femur. Defects were grafted using either HA-positive cells or heparinized bone marrow aspirate (hBMA).

Both defects grafted with MS-processed HA-positive cells and hBMA showed robust bone formation at 4 weeks, but the amount of total bone mineral in the defect was not different between the two groups as assessed by microCT. Using histomorphometric analysis, both the area of new bone formation and vascular sinusoids was significantly greater when MS processing was used.

14 Executive Summary:

This project addresses the need for improved methods of selection and concentration of connective tissue progenitors from the heterogeneous mix of cells present in a fresh bone marrow aspirate. The frequency of CTPs in bone marrow is quite low, on average 1 in

20,000 nucleated cells, and this rarity highlights the need for clinically relevant selection strategies to optimize the impact of these cells in a graft site.

This project represents the first exploration of the potential utility of using the extracellular matrix niche around a cell as a potential marker for positive or negative selection. Upon aspiration from the bone marrow, CTPs are expected to retain components of their extracellular matrix niche on their surface. We hypothesized that hyaluronan represents a distinguishing feature of the in vivo niche of one or more subsets of the heterogeneous population of CTPs in bone marrow, and would therefore provide a means for enrichment of CTPs from a fresh bone marrow aspirate. We also hypothesized that positive selection of CTPs based on hyaluronan (HA) will increase the efficacy of a bone graft in vivo due to increased concentration and prevalence of CTPs in a graft site.

The specific aims of this project were:

Specific Aim 1: To test the hypothesis that surface bound HA is a distinguishing feature that can be used for positive selection of CTPs from the heterogeneous population of cells in a fresh bone marrow aspirate. (Chapter 3)

15 Specific Aim 2: To test the hypothesis that using magnetic separation based on HA to increase the concentration and prevalence of osteogenic CTPs will increase the magnitude of new bone formation in the canine femoral multidefect model (CFMD).

• Task 1: Develop a scaled up protocol to process clinically relevant volumes of

aspirated bone marrow using a hexapole magnetic system (HMS) to increase the

concentration and prevalence of CTPs based on HA. (Chapter 5)

• Task 2: Define a method for loading HA-positive CTPs onto matrices after

magnetic separation processing. (Chapter 6)

• Task 3: Determine the effect of using magnetically separated cells from marrow

based on HA expression on the magnitude of new bone formation in vivo in the

CFMD. (Chapter 7)

In Aim 1, HA-based separation resulted in the isolation of a cell fraction enriched in highly proliferative CTPs. Human bone marrow aspirates were labeled for HA and magnetically separated into HA+++, HA+ and HA- fractions (see Chapter 3). Most nucleated cells partitioned to the HA- fraction, while the HA+++ fraction was enriched in

CTPs. The progeny of CTPs in the HA+++ fraction were significantly more proliferative than unselected CTPs from marrow. In addition, the amount of alkaline phosphatase, indicative of osteoblastic differentiation, was significantly increased in colonies formed from the HA+++ CTPs over colonies formed from the unselected marrow.

16 We hypothesized that this HA+++ cell population could provide an superior cell source for bone grafting applications, due to the increased prevalence of CTPs and the elimination of the majority of competing non-osteogenic cells in the graft site.

In order to test this hypothesis in vivo in a biologically relevant large animal model, we needed to scale up the magnetic separation system to process larger volumes of marrow

(48 cc) and to define a preferred method for loading of HA-positive cells onto scaffolds for implantation into the canine femoral multidefect model (CFMD).

Chapters 4 and 5 describe refinements to the magnetic separation protocol. A mathematical model of the magnetic separation system was developed to predict the minimum retention time required to retain positively labeled cells in the magnet. Scale up was ultimately achieved using the hexapole magnetic system, developed in the

Zborowski lab, which can process 50 cc of marrow in less than an hour. This system was used for magnetic separation based on HA for the in vivo experimentation.

Chapter 6 evaluated methods for loading of HA-positive CTPs onto osteoconductive scaffolds. A manual method of selective retention in which cells are passed through an implantable matrix, allowing CTPs to selectively attach, was evaluated in comparison to a drip soaking (DS) method. Both provided excellent CTP retention within the allograft scaffold. Drip soaking was selected for use in in vivo canine experiments due to its simplicity.

17 In chapter 7, the results of the in vivo experiment using the canine femoral multidefect

(CFMD) model were presented. The CFMD model provides rigorous comparison of bone grafting materials using four cylindrical bone defects (10 mm diameter x 15 mm long) in the canine femur. The center of the defect becomes profoundly hypoxic within 2-3 days, which enables the model to be sensitive to changes in cell transplantation strategies which limit hypoxia and improve the survival of transplanted cells.

Four defects were grafted using either HA-positive cells or heparinized bone marrow aspirate in a single animal. After 4 weeks, implants were harvested and evaluated for new bone formation using microCT and quantitative histomorphometry. Both defects grafted with MS-processed HA-positive cells and heparinized bone marrow aspirate showed robust bone formation at 4 weeks, but the amount of total bone mineral in the defect was not different between the two groups as assessed by microCT. Histomorphometric assessment was used to quantitatively compare the type and quality of the tissue formed in the defect site. Using histomorphometric analysis, both the area of new bone formation and vascular sinusoids was significantly greater when MS processing was used.

In aggregate, these data support the initial hypotheses of these projects. They provide the first proof-of-concept that HA can be used as a marker for selection of CTPs and that rapid intraoperative processing of bone marrow using HA can improve the outcome of local bone regeneration in vivo. Because the level of performance of MS processed cells and heparinized marrow delivered in a MCA scaffold left little room for improvement,

18 further assessment should be performed in a large animal model involving a larger defect that is more representative of the most challenging clinical defects.

In this project we have contributed several novel observations to the scientific community.

We have demonstrated for the first time the concept of using cell-associated HA to preferentially select CTPs. We have also shown that HA is a marker that may be used to subselect between subpopulations of CTPs, since CTPs were found in the HA- as well as the HA+ populations, but the HA+++ CTPs demonstrated greater proliferation and alkaline phosphatase expression in vitro. We have also become the first to show that the enriched

CTP population in the HA-positive fraction can be used enhance healing in a bone defect in vivo.

Two valuable tools were developed that will enable further advancement of magnetic separation processing for HA or other markers. The characterization and mathematical modeling of the commercially available EasySepTM magnetic separation system was developed to predict retention time of labeled cells in the magnet. The need for large scale separation of bone marrow resulted in the development of a completely new magnetic system for HA-based separation that can process clinically relevant volumes of marrow.

There are many directions in which this research can be expanded to explore the composition and function of HA in the CTP niche. Future work could provide clues to the strategies that CTPs use to maintain quiescence in vivo, as well as signals that provoke

19 their proliferation and differentiation. Further development of strategies for HA-based separation of CTPs may focus on refinement of the magnetic separation protocol to reduce the loss of CTPs during processing, or uncover the other cell types that co- separate with CTPs that may encourage or inhibit the colony forming efficiency of CTPs.

Alternately, development of a scaffold that presents a hylauronan binding protein can be paired with selective retention methodology to enhance CTP retention in a scaffold used for bone regeneration. This eliminates the need for magnetic separation and the associated processing time required for labeling and retention in the magnet.

20 Chapter 1:

Introduction to the Bone Tissue Engineering Paradigm,

Connective Tissue Progenitor Cells, and Hyaluronan

1.1 Clinical need for bone grafts

Bone defects, including those from severe fractures, non-unions, cavities due to trauma or tumors, segmental defects, and spinal fusion require surgical intervention and bone grafting in order to induce healing.1 Approximately 1.5 million bone grafting procedures are performed each year in the United States to promote healing in these complicated defect sites.2 Despite advances in bone grafting materials, 2-30% of these grafts ultimately fail, illustrating the need for improved technology to reduce revision surgeries and increase the rate of healing in these patients.3,4

1.2 Bone tissue engineering paradigm

Fracture healing depends on the coordinated interaction of an osteoconductive matrix, osteoinductive stimuli, and osteogenic cells. The graft site must also provide a surface for vascular tissue ingrowth and a mechanical environment conducive to bone formation.

Osteoconductive materials act as a scaffold conducive to bone growth by aiding in the ingrowth of vasculature, perisvascular tissue, and osteoprogenitor cells from intact surrounding host tissue.1 The surface chemistry, texture and three dimensional architecture of a scaffold all contribute to osteoconduction.

21 Osteoinduction refers to stimuli that promote osteogenic progenitor cells to be recruited to the graft site from surrounding host tissues.1 Osteoinductive stimuli induce osteogenic cells to activate, migrate, proliferate, and differentiate. The prototypical osteoinductive stimuli are bone morphogenetic proteins (BMPs), which induce bone formation in vivo in both orthotopic and heterotopic sites.5-8 Osteogenic cells are defined as any cell that is capable of differentiating into an osteoblast or bone forming cell.

1.3 Materials used for bone grafts

Orthopaedic surgeons have many choices of bone graft materials. The most frequently used, and the current standard of care in the field of bone grafting, is autologous cancellous bone (ACB). ACB is usually harvested from the iliac crest, and provides an osteoconductive and osteoinductive surface. ACB offers the added advantage of containing a source of autologous osteogenic cells. However, there is a limit to the amount of bone that can be harvested from a patient. In addition, there is donor site morbidity associated with ACB harvest, including scarring, blood loss, pain, nerve or arterial injury, and infection rates between 8 – 10%.9-14 These limitations underscore the need to define alternate sources of graft material that can perform as well or better than

ACB.

Allograft bone matrix is used less frequently than ACB, but its application is growing.

Allograft bone is used in approximately one third of all bone grafting procedures in the

United States.15 Like ACB, allograft also provides an osteoconductive surface, and can provide some osteoinduction.16-25 Donor allograft bone eliminates the need for a harvest site, thus avoiding the morbidity that is associated with ACB, and can be effective in

22 many types of defect sites.26 Allograft comes in many formulations including chips, powders, and fibers, and can be fitted to fill any defect geometry.

Allograft usage is limited by its increased cost and variability in the quality of allograft due to donor variability and differences in processing techniques, which affects its osteoinductive, structural, and mechanical properties.26 Despite improvement in processing methods, allograft carries at least a small risk of immune reaction and disease transmission.27,28 The use of allograft alone has an increased risk of non-union, re- fracture, and infection when compared to ACB.3, 29-34 Allograft bone implants have limited revascularization compared to ACB, partially due to the lack of donor progenitor cells in the graft site that contribute to healing in ACB grafts. In addition, the immune response that occurs during allograft incorporation retards the healing process.1 The lack of vascularization in allograft bone grafts over time may contribute to the high fracture rate, between 16-50% of bone grafts.35-36

Additional surgical options include materials pre-loaded with osteoinductive growth factors such as bone morphogenic protein 2 (BMP-2 Infuse) or osteogenic protein-1 (OP-

1 Device). BMPs were first discovered by Urist in 1965, when sites in rats implanted with demineralized bone matrix formed ectopic bone.8 The OP-1 Putty and OP-1 Implant, marketed by Stryker, are FDA-approved for revision of posterolateral lumbar spine fusion and for the treatment of long bone nonunion fractures under a Humanitarian

Device Exemption (HDE).37,38 Medtronic‟s BMP INFUSE product is FDA approved through a pre-market approval (PMA) process. It is approved clinically for oral, maxillofacial, and dental regenerative bone grafting procedures, as well as acute, open

23 fractures of the tibial shaft, and spinal fusion procedures in skeletally mature patients with degenerative disc disease.39-42

An active area of research is the development of polymer scaffolds for bone regeneration.

Polymeric scaffolds have the distinct advantages of being reproducible and stable as off- the-shelf products, but their performance alone in defect sites has yet to match the healing response provided by ACB. As a result, polymer scaffolds are commonly being developed as composite products in combination with cells and/or growth factors to boost their capacity for bone regeneration.

1.4 Need for osteogenic cells

Both allograft and polymer scaffolds can be designed to be osteoconductive, but only osteogenic cells can generate new bone to repair a bone defect. Successful bone repair or regeneration in these settings requires the addition of osteogenic connective tissue progenitors (CTP-Os).

The Muschler laboratory has used and published a nomenclature that defines osteogenic cells as a subset of the larger population of connective tissue progenitors (CTPs).43-45

CTPs are defined as tissue-resident stem or progenitor cells that proliferate to form a colony in vitro and can be induced to express one or more connective tissue phenotypes.44,45 Osteogenic CTPs include a heterogeneous mixture of stem and/or progenitors cells that are capable of osteoblastic differentiation.

24 In settings where the local population of CTPs is sufficient, they may be effectively

“targeted” using osteoconductive scaffolds or growth factors. However, in settings where the local CTP population is suboptimal, as in most complex defect sites, optimizing the bone healing response will require transplantation of CTPs from an alternative source.

Connelly and Shindell first showed the efficacy of bone marrow application in the repair and union of bony defects in humans. They injected fresh bone marrow into defects in tibial nonunions and achieved union by 6 months.46 Others have shown the advantage of adding bone marrow to nonunion defects in humans.47,48 Many preclinical studies demonstrate improved graft performance when CTPs are added, even to small graft sites in young healthy animals, supporting the premise that the CTP population is suboptimal in most clinical settings and that optimal performance from any osteoconductive or osteoinductive material may require augmentation with CTPs.45, 49-54

Osteogenic CTPs are most commonly obtained from bone marrow, often by aspiration from the iliac crest. Bone marrow aspiration causes little morbidity and provides an easily accessible source of autologous cells, offering a cell source with no immunogenic risk to the patient. The frequency of CTPs in bone marrow is quite low, on average 1 in 20,000 nucleated cells,55,56 and this rarity highlights the need for clinically relevant selection strategies to optimize the impact of these cells in a graft site.

Improved methods for CTP concentration and selection will enable surgeons to deliver more CTPs in a given site and to limit the transplantation of competing non-osteogenic cells. Implantation of fewer cells should improve CTP survival by reducing local

25 metabolic demand, which is expected to be an important variable in large bone defects and non-unions.

1.5 Osteogenic cells present in bone marrow

The existence of non-hematopoietic progenitor cells in the bone marrow was first discovered by Friedenstein in 1970.57 He cultured bone marrow cells in monolayer culture and found proliferative fibroblast-like cells that formed colonies. These colony forming units-fibroblastic (CFU-F) did not express hematopoietic markers, and the colonies were formed from a single cell.57-60

Culture conditions for these cells were further optimized, and in 1991, Arnold Caplan first introduced the term “mesenchymal stem cell” (MSC) to denote the population of adult multipotent stem cells that could differentiate down the osteogenic, chondrogenic, and adipogenic lineages in vitro.61 These cells are identified retrospectively, by their progeny after plastic adherence and colony formation in tissue culture.62,63 The MSC terminology is most often used to denote culture expanded cells that have been selected for during cell culture and comprise a homogeneous cell population. Caplan was also the first to envision the MSC as an autologous therapeutic agent in the repair of various mesenchymal tissues.61

In recent years, there has been some resistance to the use of the MSC terminology, due to the fact that continuous self-renewal is not required of a “mesenchymal stem cell”.64,65

Traditionally, an adult stem cell is required undergo asymmetrical cell division, producing a clone of itself and a daughter cell that differentiates down a lineage. The

International Society for Cellular Therapy (ISCT) issued a position statement in 2005 that

26 advocates the use of the term “multipotent mesenchymal stromal cells” (MSC) instead of mesenchymal stem cells.65,66

The definition postulated by ISCT requires multipotent mesenchymal stromal cells to satisfy 3 requirements: 65,66

1. Adherence to plastic in standard culture conditions

2. Phenotype/cell surface marker expression: Positive (>/95% positive) for CD105, CD73,

CD90 and negative (

3. In vitro differentiation to osteoblasts, adipocytes, and chondroblasts (demonstrated by staining of in vitro cell culture)

Subsets of MSC have been characterized in the last 10 years. In 2002 Catherine Verfaillie coined the term multipotent adult progenitor cells (MAPC) to denote the rare population of bone marrow derived cells within mesenchymal stem cell cultures that can replicate for more than 80 population doublings and can be induced to form endothelium and endoderm as well as the mesechymal lineages.67-70

In 2004, D‟Ippolito published a paper detailing the isolation, culture, and characterization of marrow-isolated adult multilineage inducible cells, or MIAMI cells.71 Under specific culture conditions designed to mimic the bone marrow microenvironment, MIAMI cells replicated for over 50 population doublings and, when induced, differentiated into cells from mesodermal lineages and cells with features of neuroectodermal- and endodermal- derived lineages. Bone marrow was cultured for 14 days on fibronectin at 3-5% O2 with

27 nonadherent cells, which presumably provide essential cytokines for the proliferation of the MIAMI cells.71

Other subsets of adult human bone marrow-derived MSCs have been identified, including very small embryonic-like cells (VSEL)72,73 and multipotent adult stem cells (MASC)74.

CTPs differ in morphological and surface characteristics and biological potential from preparations of these culture-expanded cells. These culture expanded cells are more homogeneous due to the selection pressures associated with the process of in vitro isolation and represent the progeny that are derived from one or more subsets of CTPs.

Cell surface markers present on native CTPs are modified after in vitro culture depending on the medium composition and culture conditions used, and several studies using flow cytometry have shown changes in cell surface marker presentation after cell culture (see

Chapter 2).75-77

1.6 Review of the properties of Hyaluronan

1.6.1 Composition

Hyaluronan (HA), also called hyaluronic acid or hyaluronate, was first discovered in

1934 by Karl Meyer and John Palmer, isolated from bovine vitreous humor and named

“hyaluronic acid” due to presence of uronic acid and its location in the hyaloid

(vitreous).78 HA is a glycosaminoglycan (GAG) present in the extracellular matrix that has several unique features. Hyaluronan is a non-sulfated, negatively charged linear molecule. It has a simple structure, consisting of repeating units of glucuronic acid and

N-acetyl-glucosamine linked by glucuronidic (1→3) bonds, and then disaccharides are

28 linked together by a hexosaminidic (1→4) bond.79 HA is the only GAG not linked to a core protein.

1.6.2. Size and physical properties

Chains of hyaluronan can be very large, reaching a length of up to 25 micrometers. The molecular weight of native HA exceeds 106 Daltons, with a range of 2000 to 25,000 disaccharide units.80 It is hygroscopic, readily attracting and retaining moisture, a feature which imparts its important role in regulating tissue hydration and osmotic balance.81 HA imparts viscoelasticity to tissues and functions as a “space saver” in the extracellular space because of its ability to occupy a large solvent domain.82

Synthesis

Chains of hyaluronan are synthesized at the plasma membrane on cytosolic side and extruded through the cell membrane to the extracellular side. This method of synthesis is unique to hyaluronan, as all other GAGs are assembled in the Golgi and subsequently transported to their final destination anchored in the cell membrane, in intracellular granules, or released into the extracellular space. Also unlike other GAGs, sugars are added to the growing HA strand at the reducing end.82

There are three known hyaluronan synthases: HAS 1-3. In vitro data suggests that HAS1 is the least active, and manufactures HA chains with a molecular weight range of 200-

2000 kDa. HAS2 is more active than HAS 1, and produces similarly sized HA molecules.

HAS2 is believed to be responsible for the increased HA deposition found during stress- induced conditions and the wound healing response. HAS3 is the most active synthase,

29 producing smaller HA fragments with a molecular weight range <100 kDa. The hyaluronan produced by HAS3 may be involved in activation of signal transduction.83

1.6.3. Degradation

Hyaluronan is constantly synthesized and degraded. By one estimate, one-third all of the

HA in the body is synthesized and degraded in a single day. 84 HA is removed by cells using receptor-mediated endocytosis and subsequent degradation in lysosomes. This can occur locally, after transport of free HA to the lymph nodes, or by endothelial cells in the liver after removal of HA from the bloodstream.82 The lining cells of the lymphatic sinuses can remove as much as 90% of the HA that enters, leaving the remaining 10% to enter the bloodstream and be removed during passage through the liver. 82

The enzymes responsible for degrading HA are a group of hyaluronidases that hydrolyze the hexosaminidic (1→4) bonds.79 Six hyaluronidases (HYALs) have been identified, but only three have demonstrated functional degradation of hyaluronan in vivo. PH-20, a hyaluronidase found on the surface of sperm, is anchored by glycosylphosphatidylinositol

(GPI) and functions to degrade the HA-rich matrix surrounding the oocyte, providing a path for the sperm to reach the oocyte and fertilize it.85 HYAL1 is present in all cells in lysosomes, and degrades HA to small fragments of 2-10 disaccharides. HYAL 1 is also present, albeit inactive, in serum in humans.86 The optimum pH for HYAL1s enzymatic activity is acidic, matching the pH of the lysosomal compartment. HYAL2 is found in the plasma membrane, anchored by GPI, and degrades HA to larger fragments around 50 disaccharides. HYAL2 degrades at a much slower rate than HYAL1.87

30 1.7. HA configuration in vivo Extracellularly, HA is found in one of three configurations: as a free-floating GAG (ex - synovial fluid, vitreous of the eye), as part of the backbone of extracellular matrix structures though proteoglycan binding, or attached to the cell surface forming pericellular coats.88

Due to its ability to act as a lubricant and hydrate tissues, free HA performs important functions in the tissue. HA is present in high concentration in the synovial fluid of the knee where HA functions to lubricate the joint, distribute load, and reduce friction. Its visco-elasticity also allows for a quick recovery after deformation of the joint during movement.82

HA is present in the extracellular matrix of many tissues, and can associate with proteoglycans to form large complexes, imparting the 3D structure of the extracellular matrix. In cartilage, aggrecan binds HA, which is stabilized by link protein, and forms large aggregates. The individual aggrecan:link protein complex requires 40 monosaccharides of HA: 10 monosaccharides for aggrecan binding, 10 for the link protein binding, and 20 monosaccharides spacing in between the associated aggrecan and link protein.89 The spacing between each aggrecan:link protein complex on n-HA has been measured as 100 monosaccharides. 89 In fact, more than 100 aggrecan and link protein complexes can bind to a single HA strand of approximately 10,000 monosaccharides.90 These massive aggregates impart visco-elasticity to cartilage. Other hyaladherins in extracellular matrices include versican, hyaluronectin, and neurocan.91-93

90 All of these matrix hyaladherins require 10-12 saccharides to bind (HA10-12).

Cells have receptors for binding HA as well. CD44 is present on most cell types, and binds HA with strong affinity. It is believed that contact between CD44 and HA affects cell functions including development, inflammation, tumor growth and metastasis.94

Internalization of HA into cells is CD44-mediated. Several variants of CD44 exist, suggesting that different variants of the receptor provide different functions for the cell.

90 CD44 only requires a hexasaccharide (HA6) sequence to bind HA.

RHAMM (receptor for hyaluronan-mediated motility expressed protein) has been shown to be an important receptor for cellular migration of several cell types. RHAMM is present on the cell surface, but is also present intracellularly; associated with the cell cytoskeleton, the mitochondria, and the nucleus.95 RHAMM on the cell surface aids in focal adhesion and migration in normal cells as well as tumor cells, when responding to oligos of HA.96,97

LYVE-1 is present on lymphatic and liver sinusoidal endothelial cells, the two main cell types responsible for removal of free HA from the lymph and the bloodstream, respectively.98,99 This receptor aids in the uptake of HA into the endothelial cells where it is degraded. The same endothelial cells that present LYVE-1 also contain a receptor for

HA known as HARE (hyaluronan receptor for endocytosis, AKA stabilin-2), whose function is unknown.100 This receptor can bind HA and contains 18-20 domains for epidermal growth factor.101

32 The pericellular coat of HA plays important roles in cell adhesion and shape, and is a dynamic process. A new pericellular coat of HA can form within 15-40 minutes with a thickness of several micrometers in cells preparing for cell division.102 Hyaluronan can be anchored to the cell surface through retention by the HAS enzyme or by binding with

CD44 and/or RHAMM receptors. At the cell surface, HA forms a pericellular matrix (aka glycocalyx) that can be visualized using a particle exclusion assay. In this assay, fixed

RBCs are added to the culture slide, and these RBCs will sediment and coat the slide everywhere except around a “clear zone” where the gel-like HA matrix excludes them.

Removal of hyaluronan with streptomyces hyaluronidase removes the pericellular coat and the RBCs are no longer excluded.102 Using this method, pericellular coats 20 microns in length have been visualized. This coincides with the average length of a straight HA molecule. The size of a cell diameter averages around 9 microns, so this coat can exceed twice the diameter of the cell. HA on the cell surface is associated with various proteoglycans whose negative charges repel each other, causing the strand of hyaluronan to stand straight out from the cell surface instead of assuming its natural random coil configuration.103 Some proteoglycans known to associate with HA on the cell surface include: aggrecan (chondrocytes), versican (smooth muscle cells, fibroblasts), neurocan (neurons), brevican (neurons), and phosphacan (neurons). 91,104-108

The addition of proteoglycans to an HA network imparts a high negative charge density, which, in turn, extends the HA chain due to the negative charge repulsion of the proteoglycan chains. Proteoglycan inclusion also increases viscosity and osmotic swelling pressure as well as stiffness of the pericellular matrix.109,110 The pore size of the

33 matrix will also decrease with increasing proteoglycans, which will directly influence the number and type of proteins that can access the immediate cell microenvironment.111

The pericellular matrix of HA and its associated molecules have functional consequences for the cell. The hexosaminidic (1→4) linkage of HA monomers is flexible and allows

HA to be “conformationally restless”, enabling the hyaluronan strand to adjust between conformations easily to fill oddly shaped voids and adjust to surfaces.112 When stretched out of its preferred configuration, HA will “push” back, an important property that aids in cell mobility during wound healing.103

Cells can also synthesize and retain HA in the form of long thick cables. This form of HA is made in reaction to inflammation, hyperglycemia, ER stress, or viral infection.113-118

These cables bind leukocytes and contain TSG-6, versican, and II heavy chains.117

Crosslinking of HA cables may promote cell receptor clustering, which in turn promotes inflammatory cell adhesion.117,119 Under conditions of ER stress, the formation of HA cables can begin intracellularly and then be exported to the extracellular side to form large crosslinked cables.115, 117, 118

Conflicting data shows the HA pericellular coat to be both anti- and pro-adhesive. During cell division, the pericellular matrix is compacted and thick. This may help the cell “ball up”, or push off the surface, and proliferate. Disruption of the pericellular matrix using

HA oligos inhibits proliferation, and cells assume a flattened morphology.102 Formation of a pericellular coat of the underside of cells may offer the cell a means to push itself off

(from a substrate or other cells) using the swelling force of the HA-proteoglycan coat.119

34 The coat also forms in areas of focal adhesion turnover and is observed along the flanks of moving cells, possibly acting as a lubricant.102,120 There is also evidence that the pericellular coat can be pro-adhesive. Obviously, HA present in cable form is pro- adhesive, indicating that the macrostructure of hyaluronan may play a role in regulating its adhesiveness.113,116,122,123

1.8 Hyaluronan in disease states: inflammation, wound healing, and angiogenesis

Native high molecular weight HA (n-HA) is anti-angiogenic, however, its degradation products (o-HA, 3-10 disaccharides) are pro-angiogenic. o-HA will stimulate endothelial cell proliferation, migration, and tube formation through interaction with CD44 and/or

RHAMM.124-129

n-HA has anti-inflammatory properties and is also anti-angiogenic, and has the ability to inhibit endothelial cell proliferation and migration, and promote cell quiescence and tissue integrity.130-133 n-HA promotes cell quiescence in multiple cell types and can cause cell cycle arrest when CD44 interacts with the intracellular protein Merlin and de- phosphorylates it.134

Inflammation induces an increase in expression of growth factors and cytokines which, in turn, increase extravasation and migration of inflammatory cells, fibroblasts, and endothelial cells to site of injury. During this time, HAS1-3 are upregulated and more n-

HA is produced at the area of tissue injury. CD44 and RHAMM are upregulated and much of the n-HA is degraded to o-HA by both HYALs and reactive oxygen and nitrogen species, which are produced in large amounts during inflammation.135-138

35 o-HA binds CD44 on macrophages which subsequently produce pro-inflammatory growth factors including tumor necrosis factor-alpha (TNF-), insulin-like growth factor

1 (IGF-1), interleukin 8 (IL-8) and interleukin 1-beta (IL1-), which will activate endothelial cells.138 In addition, o-HA can directly activate endothelial cells by binding with CD44 and RHAMM receptors present on the endothelial cell surface.139

Studies blocking RHAMM or CD44 receptors have shown the importance of the interaction with o-HA on the function of endothelial cells. Anti-RHAMM antibodies block migration of macrophages, endothelial cells, fibroblasts, and smooth muscle cells, which are all important cell types in inflammation and angiogenesis.139 In vitro, blocking of CD44 receptors inhibits endothelial cell proliferation and adhesion to HA.140

Endothelial cells require CD44 interaction with o-HA and activation of protein kinase C

(PKC) and MAP kinase before induction of cell division.126 In vivo, blocking RHAMM receptors prevented angiogenesis, demonstrating the importance of HA in this process.140

During wound healing, o-HA must be cleared from wound site and replaced with n-HA for optimal tissue remodeling to occur.

1.9 Hyaluronan in the bone marrow

The major extracellular matrix components of the bone marrow include fibronectin, collagens 1 and 4, laminin and the glycosaminoglycans heparan sulfate, chondroitin sulfate, and hyaluronan.141 Several researchers have examined histological sections of bone and shown a non-uniform distribution of HA in the bone marrow. In sections of rat tibia, strong HA staining is present in the perivascular area. HA staining was also

36 observed surrounding isolated stromal cells in the marrow space.142 Echoing these results, sections of human bone marrow show staining of HA in stromal extracellular matrix, perivascularly around the marrow sinusoids, and on the endosteal surface of bone.143-145 It has been hypothesized that HA plays a role in the stem cell niches within bone marrow.

1.10 Concept of a stem cell niche

Schofield first proposed the existence of a niche for stem cells in 1978.146 Since then, significant contributions have been made advancing this concept, and several different stem cell niches have been elucidated. Scadden described the stem cell niche as “specific anatomic locations that regulate how they (stem-cell populations) participate in tissue generation, maintenance, and repair. The niche saves stem cells from depletion, while protecting the host from over-exuberant stem-cell proliferation”.147 Importantly, the niche is functional as well as anatomic. The niche is thought to function to regulate quiescence of stem cells as well as directing their proliferation and differentiation. Some factors that can contribute to the maintenance of the niche include: the 3D architecture of the niche, cell:cell interactions or paracrine signaling, interaction with the surrounding extracellular matrix, hypoxia, ion gradients, neural inputs, and metabolic cues.147

1.11 Hematopoietic stem cell niche

For hematopoietic stem cells (HSCs), two niches within the bone marrow have been described. The first, better established, niche occurs at the endosteal surface of bone, where the most primitive HSCs exist in a specific microenvironment that maintains their quiescence and limits the HSC pool.148,149 These HSCs directly interact with the osteoblastic bone lining cells, which serve as a negative regulator of HSC proliferation. It

37 has been shown that important factors within this niche include the calcium ion gradient, necessary for homing and engraftment of HSCs into this niche using a calcium sensing receptor, and hypoxic conditions.147,150,151 Histological sections of human bone show hyaluronan expression in the endosteal area. Cell-associated HA affects adhesion, motility and growth of a wide variety of cell types.152 Association with hyaluronan may keep HSCs undifferentiated and help maintain quiescence. In vitro, binding of HSCs to

HA suppresses proliferation and differentiation.153 In the endosteal bone region, HAS3 produced HA is present in the extracellular matrix and is critical for HSC lodgment and quiescence. HAS3 knockout mice have no HA in the endosteal region. Transplanted

HSCs have decreased rates of engraftment there, and the native HSC pool is higher.147

This indicates that the presence of HA in the endosteal area aids in maintaining HSC in a resting state, regulating their numbers in that region.

The location of the second niche for HSCs is around the bone marrow microvasculature towards the center of the marrow canal, and contains the more committed, proliferating hematopoietic stem and progenitor cells.148,153 These HSCs are preparing to make mature cells that need to diapedese through vessel wall into the general circulation.147 This is also an area of strong HA staining in histological sections of human bone biopsies.

However the role of HA in this niche, if any, is unknown.

In the bone marrow, HA is synthesized by stromal cells and CD34+ hematopoietic stem cells. As the HSC becomes more committed it loses CD34 and expresses CD38 and concomitantly decreases synthesis of HA.152 However, the major producer of HA in the bone marrow is the colony-forming bone marrow-derived stromal cells. In vitro, HA

38 represents 40% of the total amount of synthesized GAGs made by passaged stromal cells.154 Alteration of culture conditions can change HA production, for example, the addition of the steroid methylprednisolone decreases amount of HA found in serum and associated with stromal cells in in vitro culture.155 HA production from bone marrow mesenchymal progenitor cells was also decreased when dexamethasone was added to culture medium.156

1.12 Evidence for CTP niche

Evidence for a third stem cell niche in bone marrow is accumulating. Due to the relative infrequency of CTPs, many investigators have focused on phenotyping the culture expanded progeny of CTPs, termed mesenchymal stem cells. This lack of characterization of the in vivo phenotype means their niche hasn‟t yet been fully explored.

However, there is increasing evidence that these cells may occupy a perivascular niche in vivo.

Using combinations of cell surface markers to identify mesenchymal stem cells (MSCs), immunohistochemical staining of CD45-/CD31-/SCA-1+/THY-1+ cells showed localization of these cells in perivascular sites.157 In another study, Stro-1+/CD146+ cells were found lining blood vessels in human bone marrow and dental pulp.158 Stro-1 has been used to isolate colony forming MSCs, while CD146 is a pericyte marker. Cultured

MSCs express CD146, and lends credence to the hypothesis that MSCs may have a pericyte origin.159

Sacchetti et al. separated CD45- CD146+ and CD45- CD146- cells from bone marrow aspirates and found all colony forming cells in the CD45- CD146+ fraction. The CD45-

39 CD146+ cells also expressed alkaline phosphatase and CD105.160 Crisan et al. performed fluorescence activated cell sorting on cells dissociated from multiple tissues, including bone marrow. Cells were first gated for CD56- to remove contaminating cells with myogenic potential, then CD45- to eliminate all hematopoietic cells, and finally selected for CD146+ CD34- cells. These CD56- CD45- CD146+ CD34- cells differentiated along the chondrogenic, adipogenic, and osteogenic lineages in vitro, and formed bony nodules in vivo when implanted into an immunocompromised mouse model.161 HA is present around the marrow sinusoids, and could be a possible component of the CTP niche.

The perivascular niche would provide CTPs easy access to blood vessels for homing to injured tissues, and would provide a niche for these stem cells in many tissues. MSCs can be isolated based on plastic adherence and colony formation in in vitro culture. Samples of tissues throughout the body isolated similar populations of MSCs in all tissues.162

In fact, some researchers suggest that the in vivo identity of MSCs may be pericytes.163,164

Pericytes line the abluminal side of vasculature and are found in tissues throughout the body. Pericytes are extensively branched cells that envelope endothelial cells and wrap around the vessel.165 Pericytes fuction to stabilize the vessel, regulate vascular tone, and synthesize matrix proteins, among others.165 There are subsets of pericytes in bone marrow including adventitial reticular cells and Westen- Bainton cells, all of which may act as MSCs in vivo.

40 In the arterial and capillary areas of BM, pericytes have been shown to express both the pre-osteoblastic marker alkaline phosphatase (AP) and alpha smooth muscle actin

(aSMA). In the venous portion, adventitial reticular cells express AP but not aSMA.166

Adventitial reticular cells are specialized pericytes that have long processes reaching from the sinus wall into the adjacent hematopoietic marrow space. These processes are responsible for establishing cell:cell contacts that convey cues to maturing blood cells. 166

Westen- Bainton cells are another possible CTP source. These are fibroblast-like cells that express AP and are associated with the outer surface of marrow sinusoids.167-170

There is evidence that CTPs can be found in multiple tissue compartments in bone. CTPs are found in bone marrow, in the periosteum, surrounding the vasculature as pericytes or reticular cells, and on the trabecular surface of bone. At least some CTPs in native tissues are multipotent (capable of forming bone, cartilage, fat, muscle, fibrous tissue, and stroma).171-172 CTPs in the outer periosteum have both osteogenic and chondrogenic potential.173-175 Multipotent CTPs are present on the trabecular surface of bone.55,176-179

Jones et al. recently showed a 65-fold increase of CFU in the CD45LOWCD271+ fraction isolated by enzymatic digestion from trabecular bone over bone marrow aspiration alone.179 This indicates that CTPs may be present in much higher frequencies in bone marrow than can be assessed by using a bone marrow aspirate alone.

A definitive surface marker profile has yet to be established for the CTP population in native tissues, including bone marrow, making identification of these cells in vivo or using rapid isolation a continuing challenge. Upon aspiration or mechanical extraction of cells from the bone marrow, CTPs are expected to retain components of their

41 extracellular matrix niche on their surface. We hypothesize that HA may represent a distinguishing feature of at least one of the in vivo niches occupied by CTPs, and therefore could provide a means for enrichment of CTPs from the heterogeneous population of cells in a fresh bone marrow aspirate. The concept of using retained HA from the cell‟s in vivo niche to preferentially select CTPs has not been previously investigated. Additionally, we hypothesize that positive selection of CTPs based on hyaluronan will increase the efficacy of a bone graft in vivo due to increased concentration and prevalence of CTPs in a graft site.

The specific aims of this project are:

Specific Aim 2: To test the hypothesis that using magnetic separation based on HA to increase the concentration and prevalence of osteogenic CTPs will increase the magnitude of new bone formation in the canine femoral multidefect model (CFMD).

• Task 1: Develop a scaled up protocol to process clinically relevant volumes of

aspirated bone marrow using a hexapole magnetic system (HMS) to increase the

concentration and prevalence of CTPs based on HA. (Chapter 5)

• Task 2: Define a method for loading HA-positive CTPs onto matrices after

magnetic separation processing. (Chapter 6)

42 • Task 3: Determine the effect of using magnetically separated cells from marrow

based on HA expression on the magnitude of new bone formation in vivo in the

CFMD. (Chapter 7)

1.13 References

1. Rockwood and Green 6th Ed. “Fractures in Adults” Editors R. Bucholz, J Heckman, C.

Court-Brown. Lippincott, Williams, & Wilkins 2006. Philadelpia PA

2. Bishop G.B., Einhorn T.A. Current and future clinical applications of bone

morphogenic proteins in orthopaedic trauma surgery. Internal Orthpaedics 2007.

31:721-727

3. Berrey B.H. Jr, Lord C.F., Gebhardt M.C., Mankin H.J. Fractures of allografts.

Frequency, treatment, and end-results. J Bone Joint Surg Am. 1990. 72(6):825-33

4. Bach, A.W., Hansen S.T. Delayed union, nonunion, and malunion of the tibial shaft.

Surgery of the Musculoskeletal System, ed. C.M. Evarts. Vol. 3. 1983, New York,

NY: Churchill Livingstone

5. Buring K, On the origin of cells in heterotopic bone formation. Clin Orthop Relat Res.

1975 (110):293-301.

6. Ohgushi, H., V.M. Goldberg, and A.I. Caplan, Heterotopic osteogenesis in porous

ceramics induced by marrow cells. J Orthop Res, 1989. 7(4): p. 568-78.

7. Urist, M.R., R.J. DeLange, and G.A. Finerman, Bone cell differentiation and growth

factors. Science, 1983. 220(4598): p. 680-6.

8. Urist, M.R., Bone: formation by autoinduction. Science, 1965. 150(698): p. 893-9.

43 9. Younger, E.M., Chapman M.W. Morbidity at bone graft donor sites. J Orthop

Trauma, 1989. 3(3):192-195

10. Arrington, E.D., et al., Complications of iliac crest bone graft harvesting. Clin Orthop

Relat Res, 1996. 329:300-309

11. St John, T.A., et al. Physical and monetary costs associated with autogenous bone

graft harvesting. Am J Orthop, 2003. 32(1):18-23

12. Banwart J.C., Asher M.A., Hassanein R.S. Iliac crest bone graft harvest donor site

morbidity: a statistical evaluation. Spine 1995. 20:1055-1060.

13. Fowler BL, Dall BE, Rowe DE. Complications associated with harvesting

autogenous iliac bone graft Am J Orthop 1995 24:895-903.

14. Goulet JA, Senunas LE, DeSilva GL. Et al Autogenous iliac crest bone graft:

complications and functional assessment. Clin Orthop Relat Res 1997 339:76-81.

15. Boyce T, Edwards J, Scarborough N. Allograft bone: the influence of processing on

safety and performance. Orthop Clin North Am 1999 30:571-581.

16. Bolander, M.E. and G. Balian, The use of demineralized bone matrix in the repair of

segmental defects. Augmentation with extracted matrix proteins and a comparison

with autologous grafts. J Bone Joint Surg Am, 1986. 68(8): p. 1264-74.

17. Einhorn, T.A., et al., The healing of segmental bone defects induced by demineralized

bone matrix. A radiographic and biomechanical study. J Bone Joint Surg Am, 1984.

66(2): p. 274-9.

18. Glowacki, J. and J.B. Mulliken, Demineralized bone implants. Clin Plast Surg, 1985.

12(2): p. 233-41.

44 19. Harakas, N.K., Demineralized bone-matrix-induced osteogenesis. Clin Orthop, 1984.

188: p. 239-51.

20. Muthukumaran, N. and A.H. Reddi, Bone matrix-induced local bone induction. Clin

Orthop, 1985. 200:p. 159-64.

21. Nade, S. and R.G. Burwell, Decalcified bone as a substrate for osteogenesis. An

appraisal of the interrelation of bone and marrow in combined grafts. J Bone Joint

Surg Am, 1977. 59: p. 189-96.

22. Senn, S., On the healing of aseptic cavities by implantation of antiseptic decalcified

bone. Am J Med Sci, 1989. 98(3): p. 219.

23. Tuli, S.M. and A.D. Singh, The osteoninductive property of decalcified bone matrix.

An experimental study. J Bone Joint Surg Br, 1978. 60(1): p. 116-23.

24. Bauer, T.W. and G.F. Muschler, Bone graft materials. An overview of the basic

science. Clin Orthop, 2000. 371: p. 10-27.

25. Aurori, B.F., et al., Pseudarthrosis after spinal fusion for scoliosis. A comparison of

autogeneic and allogeneic bone grafts. Clin Orthop, 1985. 199: p. 153-8.

26. Nandi, S.K., Roy S., Mukherjee P., Kundu B., De D.K., & Basu D. Orthopaedic

applications of bone graft & graft substitutes: a review. Indian J Med Res 2010.

132:15-30

27. Parikh SN. Bone graft substitutes: past, present, and future. J Postgrad Med 2002

48:142-148.

28. Fleming JE, Cornell CN, Muschler GF. Bone cells and matrices in orthopaedic tissue

engineering. Orthop Clin North Am 2000 31:357-374.

45 29. Burton, C.V., et al., Causes of failure of surgery on the lumbar spine. Clin Orthop,

1981. 157: p. 191-9.

30. Friedlaender, G.E., D.M. Strong, and K.W. Sell, Studies on the antigenicity of bone.

II. Donor-specificanti-HLA antibodies in human recipients of freeze-dried allografts.

J Bone Joint Surg Am, 1984. 66(1):p. 107-12.

31. Stabler, C.L., et al., Failure of posterior cervical fusions using cadaveric bone graft

in children. J Bone Joint Surg Am, 1985. 67(3): p. 371-5.

32. Tomford, W.W., R.J. Starkweather, and M.H. Goldman, A study of the clinical

incidence of infection in the use of banked allograft bone. J Bone Joint Surg Am,

1981. 63(2): p. 244-8.

33. Wheeler DL, Enneking WF Allograft bone decreases in strength in vivo over time.

Clin Orthop Relat Res. 2005 (435):36-42.

34. Betz RR. Limitations of autograft and allograft: new synthetic solutions.

Orthopaedics. 2002 25(5 suppl):561-570.

35. Enneking WF, Mindell ER. Observations on massive retrieved human allograft. J

Bone Joint Surg Am 1991 73A:1123-1142.

36. Thompson RC Jr, Pickvance EA, Garry D. Fractures in large segment allografts. J

Bone Joint Surg Am 1993 75A:1663-1673.

37. Friedlaender, G.E., et al., Osteogenic protein-1 (bone morphogenetic protein-7) in the

treatment of tibial nonunions. J Bone Joint Surg Am, 2001. 83-A Suppl 1(Pt 2): p.

S151-8.

46 38. FDA Summary of Safety and Probable Benefit Document. Located at:

http://www.fda.gov/medicaldevices/productsandmedicalprocedures/deviceapprovalsa

ndclearances/recently-approveddevices/ucm081181.htm

39. FDA Summary of Safety and Probable Benefit Document. Located at:

http://www.fda.gov/MedicalDevices/ProductsandMedicalProcedures/DeviceApproval

sandClearances/Recently-ApprovedDevices/ucm083423.htm

40. Burkus, J.K., et al., Is INFUSE bone graft superior to autograft bone? An integrated

analysis of clinical trials using the LT-CAGE lumbar tapered fusion device. J Spinal

Disord Tech, 2003. 16(2): p. 113-22.

41. Einhorn, T.A., Clinical applications of recombinant human BMPs: early experience

and future development. J Bone Joint Surg Am, 2003. 85-A Suppl 3: p. 82-8.

42. Govender, S., et al., Recombinant human bone morphogenetic protein-2 for treatment

of open tibial fractures: a prospective, controlled, randomized study of four hundred

and fifty patients. J Bone Joint Surg Am, 2002. 84-A(12): p. 2123-34.

43. Muschler, G.F., R.J. Midura, and C. Nakamoto, Practical Modeling Concepts for

Connective Tissue Stem Cell and Progenitor Compartment Kinetics. J Biomed

Biotechnol, 2003 (3): p. 170-193.

44. Muschler G., Midura R. Connective tissue progenitors: practical concepts for clinical

applications. Clin Orthop 2002. 395:66-80

45. Muschler G, Nakamoto C, Griffith L. Engineering principles of clinical cell-based

tissue engineering. J Bone Joint Surg 2004. 86A(7):1541-1558

46. Connolly, J., et al., Development of an osteogenic bone-marrow preparation. J Bone

Joint Surg Am, 1989. 71(5): p. 684-91.

47 47. Garg, N.K., S. Gaur, and S. Sharma, Percutaneous autogenous bone marrow grafting

in 20 cases of ununited fracture. Acta Orthop Scand, 1993. 64(6): p. 671-2.

48. Healey, J.H., et al., Percutaneous bone marrow grafting of delayed union and

nonunion in cancer patients. Clin Orthop, 1990. 256: p. 280-5.

49. Hernigou, P., Poignard A, Beaujean F, Rouard H. Percutaneous autologous bone-

marrow grafting for nonunions. Influence of the number and concentration of

progenitor cells, J Bone Joint Surg Am, 2005. 87 (7):1430-7

50. Muschler, G. F., Nitto H., Matsukura Y, Boehm C., Valdevit A., Kambic H.E.,

Davros W.J., Powell K. & Easley K. Spine fusion using cell matrix composites

enriched in bone marrow-derived cells, Clin Orthop Rel Res, 2003. 407:102-18

51. Muschler, G. F., Matsukura Y., Nitto H., Boehm C.A., Valdevit A.D., Kambic H.E.,

Davros W.J., Easley K.A. & Powell K., Selective retention of bone marrow-derived

cells to enhance spinal fusion, Clin Orthop, 2005. (432):242-51

52. Tiedeman, J. J., Connolly J.F., Strates B.S. and Lippiello L., Treatment of nonunion

by percutaneous injection of bone marrow and demineralized bone matrix. An

experimental study in dogs, Clin Orthop, 1991. 268:294-302

53. Connolly, J. F., Injectable bone marrow preparations to stimulate osteogenic repair,

Clin Orthop, 1995. 313:8-18.

54. Hernigou, P., Poignard A., Manicom O., Mathieu G., & Rouard H. The use of

percutaneous autologous bone marrow transplantation in nonunion and avascular

necrosis of bone, J Bone Joint Surg Br, 2005. 87 (7):896-902

48 55. Muschler G., Boehm C., Easley K. Aspiration to obtain osteoblast progenitor cells

from human bone marrow: the influence of aspiration volume. J Bone Joint Surg Am

1997. 79:1699-709

56. Majors A.K., Boehm C.A., Nitto H., Midura R.J., Muschler G.F. Characterization of

human bone marrow stromal cells with respect to osteoblastic differentiation. J

Orthop Res 1997. 15:546-57

57. Friendenstein AJ, Chalakhjan RK and Lalykina KS. The Development of Fibroblast

Colonies in Monolayer Cultures of Guinea Pig Bone Marrow and Spleen Cells. Cell

Tissue Kinetics. 1970; 3:393-403.

58. Friendenstein AJ, Deriglasova UF, Kulagina NN, et al. Precursors for Fibroblasts in

Different Populations of Hematopoietic Cells as Detected by the In Vitro Colony

Assay Method. Experimental Hematology. 1974; 2:83-92.

59. Friedenstein AJ. Stromal Mechanisms of Bone Marrow: Cloning In Vitro and

Retransplantation In Vivo. In: Thienfelder S (ed) Immunobiology of Bone Marrow

Transplantation. Springer-Verlag. Berlin. 1980. p19-29.

60. Friendenstein AJ, Chalakhjan RK, and Gerasimov UV. Bone Marrow Osteogenic

Stem Cells: In vitro Cultivation and Transplantation in Diffusion Chambers. Cell

Tissue Kinetics. 1987; 3:393-402.

61. Caplan, A. Mesenchymal Stem Cells JOR 1991 641-650 Vol 9

62. Pittenger MF, Mackay AM, Beck SC, et al. Multilineage potential of adult human

mesenchymal stem cells. Science 1999 284:143–7.

63. Jones E, Yang X. Mesenchymal stem cells and bone regeneration: Current status

Injury, 2011 42:562–568

49 64. Bianco P, Robey PG, Simmons PJ. Mesenchymal stem cells: revisiting history,

concepts, and assays. Cell Stem Cell 2008;2:313–9.

65. Dominici M, Le Blanc K, Mueller I, Horwitz E et al. Minimal criteria for defining

multipotent mesenchymal stromal cells. The International Society for Cellular

Therapy position statement Cytotherapy 2006 8(4):315-317

66. Horwitz E, Le Blanc K, Dominici M et al. Clarification of the nomenclature for

MSC: the International Society for Cellular Therapy position statement. Cytotherapy

2005 7:393- 5.

67. Jiang Y, Jahagirdar B, Reinhart R, et al. Pluripotency of mesenchymal stem cells

derived from adult marrow nature 2002 418: 41-49

68. Reyes, M. et al. Purification and ex vivo expansion of postnatal human marrow

mesodermal progenitor cells. Blood 2001 9: 2615–2625.

69. Reyes, M. et al. Origin of endothelial progenitors in human post-natal bone marrow.

J. Clin. Invest. 2002 109:337–346.

70. Schwartz, R. E. et al. Multipotent adult progenitor cells from bone marrow

differentiate into functional hepatocyte-like cells. J. Clin. Invest. 2002 109:1291–

1302.

71. D‟Ippolito G, Diabira S, Howard G, Schiller P et al. Marrow-isolated adult

multilineage inducible (MIAMI) cells, a unique population of postnatal young and old

human cells with extensive expansion and differentiation potential Journal of Cell

Science 2004 117: 2971-2981

72. Kucia M, Halasa M, Wysoczynski M, et al. Morphological and molecular

characterization of novel population of CXCR4+ SSEA-4+ Oct-4+ very small

50 embryonic-like cells purified from human cord blood: preliminary report. Leukemia

2007 21:297–303.

73. Kucia M, Reca R, Campbell FR, et al. A population of very small embryonic-like

(VSEL) CXCR4(+)SSEA-1(+)Oct-4+ stem cells identified in adult bone marrow.

Leukemia 2006 20:857–69.

74. Beltrami AP, Cesselli D, Bergamin N, et al. Multipotent cells can be generated in

vitro from several adult human organs (heart, liver, and bone marrow). Blood 2007

110:3438–46.

75. Battula VL, Treml S, Bareiss PM, et al. Isolation of functionally distinct mesenchymal

stem cell subsets using antibodies against CD56, CD271, and mesenchymal stem cell

antigen-1. Haematologica 2009;94:173–84.

76. Vogel W, Grunebach F, Messam CA, et al. Heterogeneity among human bone

marrow-derived mesenchymal stem cells and neural progenitor cells. Haematologica

2003;88:126–33

77. Deschaseaux F, Pontikoglou C, Sensébé L Bone regeneration: the stem/progenitor

cells point of view J. Cell. Mol. Med. 2010 14:103-115

78. Laurent TC. 2002 “The Tree”: Hyaluronan Research in the 20th Century. Hyaluronan

Today Glycoforum article:

http://www.glycoforum.gr.jp/science/hyaluronan/HA23/HA23E.html

79. Jiang D, Liang J, and Noble P. Hyaluronan in tissue injury and repair Annu Rev Cell

Dev Biol 2007 23:435-61

80. Toole B. Hyaluronan: from extracellular glue to pericellular clue. Nat Rev Cancer

2004 4:528-539

51 81. Tammi M, Day A, and Turley E. Hyaluronan and homeostasis: a balancing act. J

Biol Chem 2002 277:4581-4584

82. Fraser J, Laurent T, and Laurent U. Hyaluronan: its nature, distribution, functions

and turnover. J of Internal Medicine 1997 242:27-33

83. Stern R. Hyaluronan metabolism: a major paradox in cancer biology. Pathol Biol

(Paris) 2005 53:372-382

84. Stern R. Hyaluronan catabolism: a new metabolic pathway. Eur J Cell Biol 2004

83(7): 317–25

85. Cherr G, Meyers S, Yudin A, Vandevoort C, Myles D et al. The PH-20 protein in

cynomolgus macaque spermatozoa: identification of two different forms exhibiting

hyaluronidase activity. Dev Biol 1996. 175:142-53

86. Csoka A, Frost G, and Stern R. The six hyaluronidase-like genes in the human and

mouse genomes. Matrix Biol 2001 20:499-508

87. Lepperdinger G, Mullegger J, and Kreil G. Hyal2-less active, but more versatile?

Matrix Biol 2001 20:509-514

88. Wang A and Hascall V. Hyaluronan structures synthesized by rat mesangial cells in

response to hyperglycemia induce monocyte binding. J Biol Chem 2004 279:10279-

10285

89. Faltz L, Caputo C, Kimura J, Schrode J, and Hascall V Structure of the complex

between hyaluronic acid, the hyaluronic acid binding region, and the link protein of

proteoglycan aggregates from the swarm rat chondrosarcoma. J Biol Chem 1979

254(4):1381-1387

52 90. Knudson C, Knudson W. Hyaluronan-binding proteins in development, tissue

homeostasis, and disease. FASEB 1993 7:1233-1241

91. LeBaron R, Zimmerman D, and Ruoslahti E. Hyaluronate binding properties of

versican. J Biol Chem 1992 267:10003-10010

92. Delpech B, Delpech A, Bruckner G, Girard N, and Maingonnat C. Hyaluronan and

hyaluronectin in the nervous system The Biology of Hyaluronan, Ciba Foundation

Symposium 1989 Evered D, Whelan J, eds. Wiley Press 208-232

93. Rauch U, Karthikeyan L, Maurel P, Margolis RU and Margolis RK. Cloning and

primary sequence of neurocan, a developmentally regulated, aggregating chondroitin

sulfate proteoglycan of brain. J Biol Chem 1992 267:19536-19547

94. Lesley J, Hyman R, and Kincade P. CD44 and its interaction with extracellular

matrix Adv Immunol 1993 54:271-335

95. Turley E, Noble P, and Bourguignon L. Signaling properties of hyaluronan receptors.

J Biol Chem 2002 277:4589-4592

96. Hall C, Lange L, Prober D, Zhang S, and Turley E. pp60(c-src) is required for cell

locomotion regulated by the hyaluronan receptor RHAMM. Oncogene 1996 13:2213-

2224

97. Zhang S, Chang M, Zylka D, Turley E et al. The hyaluronan receptor RHAMM

regulates extracellular-regulated kinase. J Biol Chem 1998 273:11342-11348

98. Banerji S, Ni J, Wang S, Clasper S, Su J, et al Lyve-1, a new homologue of the CD44

glycoprotein, is a lymph-specific reseptor for hyaluronan. J Cell Biol 1999 144:789-

801

53 99. Mouta Carreira , Nasser S, di Tomaso E, Padera T, Boucher Y et al LYVE-1 is not

restricted to the lymph vessels: expression in normal liver blood sinusoids and down-

regulation in human liver cancer and cirrhosis. Cancer Res 2001 61:8079-8084

100. Zhou B, Weigel J, Fauss L, and Weigel P. Identification of the hyaluronan

receptor for endocytosis (HARE) J Biol Chem 2000 275:37733-37741

101. Politz O, Gratchev A, McCourt P, Schledzewski K, Guillot P et al. Stabilin-1 and

-2 constitute a novel family of fasciclin-like hyaluronan receptor homologues

Biochem J 2002 362:155-164

102. Evanko S, Angello J, and Wight T. Formation of hyaluronan and versican rich

pericellular matrix is required for proliferation and migration of vascular smooth

muscle cells. Arterioscler Thromb Vasc Biol 1999 19:1004-1013

103. Evanko S, Tammi M, Tammi R, and Wight T. Hyaluronan-dependent pericellular

matrix. Adv Drug Deliv Rev 2007 59:1351-1365

104. Knudson C Hyaluronan receptor-directed assembly of chondrocyte pericellular

matrix. J Cell Biol 1993 120:825-834

105. Wight T, Potter-Perigo S, and Aulinskas T. Proteoglycans and vascular cell

proliferation. Am Rev Respir Dis 1989 140:1132-1135

106. Yamaguchi Y. Lecticans: organizers of the brain extracellular matrix Cell Mol

Life Sci 2000 57:276-289

107. Carulli D, Rhodes K, Brown D, Bonnert T, Fawcett J et al. Composition of

perineuronal nets in the adult rat cerebellum and the cellular origin of their

components. J Comp Neurol 2006 494:559-577

54 108. Deepa S, Carulli D, Galtrey C, Rhodes K, Fawcett J et al. Composition of

perineuronal nets extracellular matrix in the rat brain: a different disaccharide

composition for the net-associated proteoglycans. J Biol Chem 2006 281:17789-

17800

109. Mow V, Mak A, Lai W, Rosenberg L, and Tang L. Viscoelastic properties of

proteoglycan subunits and aggregates in varying solution concentrations. J

Biomechanics 1984 17:325-338

110. Morgelin M, Paulsson M, Heinegard D, Aebi U, and Engel J. Evidence of a

defined spatial arrangement of hyaluronate in the central filament of cartilage

proteoglycan aggregates. Biochem J 1995 307:595-601

111. Heldin P, Suzuki M, Teder P, Pertoft H. Chondroitin sulfate proteoglycan

modulates the permeability of hyaluronan containing coat around normal human

meseothelial cells. J Cell Physiol 1995 165:54-61

112. Sheehan J, Almond A. Hyaluronan: static, hydrodynamic and molecular dynamic

views Nov 1, 2001. Glycoforum:

http://www.glycoforum.gr.jp/science/hyaluronan/hyaluronanE.html

113. de la Motte C, Hascall V, Drazba J, Bandyopadhyay S, and Strong S.

Mononuclear leukocytes bind to specific hyaluronan structures on colon mucosal

smooth muscle cells treated with polyinosinic acid:polycytidylic acid: inter-alpha-

trypsin inhibitor is crucial to structure and function. Am J Pathol 2003 163:121-133

114. Lesley J, Gal I, Mahoney D, Cordell M, Rugg M, Hyman R, Day A, Mikecz K.

TSG-6 modulates the interaction between hyaluronan and cell surface CD44. J Biol

Chem 2004 279:25745-25754

55 115. Majors A, Austin R, de la Motte C, Pyeritz R, Hascall V, Kessler S, Sen G, and

Strong S. Endoplasmic reticulum stress induces hyaluronan deposition and leukocyte

adhesion. J Biol Chem 2003 278:47223-47231

116. Selbi W, de la Motte C, Hascall V, Day A, Bowen T, Phillips A. Characterization

of hyaluronan cable structure and function in renal proximal tubular epithelial cells.

Kidney Int 2006 70:1287-1295

117. Hascall V, Majors A, de la Motte C, Evanko S, Wang A, Drazba J,Strong S, and

Wight T. Intracellular hyaluronan: a new frontier for inflammation? Biochim

Biophys Acta 2004 1673:3-12

118. Ren J, Hascall V, and Wang A. Cyclin D3 mediates synthesis of a HA matrix that

is adhesive for monocytes in mesangial cells stimulated to divide in hyperglycemic

medium J Biol Chem 2009 284:16621-16632

119. Day A, de la Motte C. Hyaluronan crosslinking: a protective mechanism in

inflammation? Trends Immunol 2005 26:637-643

120. Yamagata M, Saga S, Kato M, Bernfield M, and Kimata K. Selective distributions

of proteoglyxans and their liogands in pericelluar matrix of cultured fibroblasts.

Implications for their roles in cell-substratum adhesion. J Cell Sci 1993 106:55-65

121. Turley E and Torrance J. Localization of hyaluronate and hyaluronate-binding

protein on motile and non-motile fibrobasts. Exp Cell Res 1984 161:17-28

122. Wilkinson T, Bressler S, Evanko S, Braun K, and Wight T. Overexpression of

hyaluronan synthases alters vascular smooth muscle cell phenotype and promotes

monocyte adhesion. J Cell Physiol 2006 206:378-385

56 123. Wang A, Hascall V. Hyaluronan structures synthesized by rat mesangial cells in

response to hyperglycemia induce monocyte adhesion J Biol Chem 2004 279:10279-

10285

124. West D, and Kumar S. The effect of hyaluronate and its oligosaccharides on

endothelial cell proliferation and monolayer integrity. Exp Cell Res 1989 183:179-

196

125. Deed R, Rooney P, and Kumar S et al. Early response gene signaling is induced

by angiogenic oligosaccharides of hyaluronan in endothelial cells. Inhibition by non-

angiogenic, high molecular weight hyaluronan. Int J Cancer 1997 71:116-122

126. Slevin M, Krupinski J, Kumar S, and Gaffney J. Angiogenic oligosaccharides of

hyaluronan induce protein tyrosine kinase activity in endothelial cells and activate a

cytoplasmic signal transduction pathway resulting in proliferation. Lab Invest 1998

78:987-1003

127. Slevin M, Kumar S, and Gaffney J. Angiogenic oligosaccharides of hyaluronan

induce multiple signaling pathways affecting vascular endothelial cell mitogenic and

wound healing responses. J Biol Chem. 2002 43:41046-41059

128. Slevin M, West D, and Kumar S. Hyaluronan, angiogenesis and malignant

disease. Int J Cancer 2004 109:793-794

129. Sattar A, Rooney P, Kumar S, and Ledger P. Application of angiogenic

oligosaccharides of hyaluronan increases blood vessel numbers in rat skin J Invest

Dermatol 1994 103:576-579

130. Rooney P, Kumar S, Ponting J, and Wang M. The role of hyaluronan in tumor

neovascularization. Int J cancer 1995 60:632-636

57 131. West D and Fan T. Hyaluronan as a regulator of angiogenesis in wound healing.

Fan T-PD, Kohn E.C (Eds), The New Angiotherapy. Humana Press Ch11, 2002 177-

188

132. West D, Shaw D, Longaker M, et al. Fibrotic healing of adult and late gestation

fetal wounds correlates with increased hyaluronidase activity and removal of

hyaluronan. Int J Biochem Cell Biol 1997 29:201-210

133. Mackool R, Gittes G, and Longaker M. Scarless healing. The fetal wound. Clin

Plast Surg 1998 25:357-365

134. Morrison H, Sherman L, Herrlich P et al. The NF2 tumor suppressor gene product,

merlin, mediates contact inhibition of growth through interactions with CD44. Genes

Dev 2001 15:968-980

135. Lovvorn H, Cass D, Savani R et al. Hyaluronan receptor expression increases in

fetal excisional skin wounds and correlates with fibroplasias. J Pediatr Surg 1998

33:1062-1070

136. Taylor K, Trowbridge J, Simon J, Gallo R et al. Hyaluronate fragments stimulate

endothelial recognition of injury through TLR4. J Biol Chem 2004 279:17079-17084

137. Tammi R, Pasonen-Seppanen S, Tammi M et al. Hyaluronan synthase induction

and hyaluronan accumulation in mouse epidermis following skin injury J Invest

Dermatol 2005 124:898-905

138. Noble P. Hyaluronan and its catabolic products in tissue injury and repair.

Matrix Biol 2002 21:25-29

58 139. Slevin M, Krupinski J, Gaffney J, Kumar S et al. Hyaluronan-mediated

angiogenesis in vascular disease: uncovering RHAMM and CD44 receptor signaling

pathways Matrix Biol 2007 26:58-68

140. Savani R, Cao G, Pooler P, DeLisser H et al. Differential involvement of the

hyaluronan (HA) receptors CD44 and receptor for HA-mediated motility in

endothelial cell function and angiogenesis. J Biol Chem 2001 276:36770-36778

141. Masellis A. 2003 Hyaluronan and hyaluronan synthesis in the bone marrow

microenvironment: a comparison of patients with multiple myeloma and healthy

donors. Hyaluronan Today Glycoforum

http://www.glycoforum.gr.jp/science/hyaluronan/hyaluronanE.html

142. Midura R, Su S, Morcuende J et al. Parathyroid hormone rapidly stimulates

hyaluronan synthesis by periosteal osteoblasts in the tibial diaphysis of the growing

rat. J Biol Chem 2003. 278:51462-51468

143. Avidgor A, Goichberg P, Shivtiel S et al. CD44 and hyaluronic acid cooperate

with SDF-1 in the trafficking of human stem/progenitor cells to bone marrow. Blood

2004. 103(8):2981-2989

144. Sundstrom G, Lofvenberg E, Hassan I et al. Localisation and distribution of

hyaluronan in normal bone marrow matrix: a novel method to evaluate impending

fibrosis? Eur J Haematol 2002. 68:194-202

145. Sundstrom G, Dahl I, Hultdin M et al. Bone marrow hyaluronan distribution in

patients with acute myeloid leukemia Medical Oncology 2005. 22:71-78

146. Schofield, R. The relationship between the spleen colony-forming cell and the

haemopoietic stem cell. Blood Cells 1978;4:7–25

59 147. Scadden D. 2006 The stem cell niche as an entity of action Nature 441:1075-1079

148. Nilsson S, Johnston H, and Coverdale J. Spatial localization of transplanted

hemopoietic stem cells: inferences for the localization of stem cell niches. Blood 2001

97:2293-2299

149. Nilsson S, Haylock D, Johnston H et al. Hyaluronan is synthesized by primitive

hematopoietic cells, participates in their lodgment at the endosteum following

transplantation, and is involved in the regulation of their proliferation and

differentiation in vitro. Blood 2003 101(3):856-862

150. Moore K and Lemischka I. Stem cells and their niches. Science 2006 311:1880-

1885

151. Ballen K Targeting the stem cell niche: squeezing blood from bones. Bone

Marrow Transplantation 2007 1-7

152. Haylock D and Nilsson S. The role of hyaluronic acid in hematopoietic stem cell

biology. Regenerative Medicine 2006 1(4):437-445

153. Lord B. The architecture of bone marrow cell populations. International J of Cell

Cloning 1990 8:317-331

154. Wight T, Kinsella M, and Keating A. Proteoglycans in human long-term bone

marrow cultures: biochemical and ultrastructural analyses. Blood 1986 67(5):1333-

1343

155. Siczkowski M, Andrew T, and Amos S. Hyaluronic acid regulate the functions

and distribution of sulfated glycosaminoglycans in bone marrow stromal cultures.

Exp Hematol 1993 21(1):126-130

60 156. Calabro T, Oken M, Hascall V, and Masellis A. Characterization of hyaluronan

synthase expression and hyaluronan synthesis in bone marrow mesenchymal

progenitor cells: predominant expression of HAS1 mRNA and up-regulated

hyaluronan synthesis in bone marrow cells derived from multiple myeloma patients

Blood 2002 100:2578-2585

157. Blashki D, Short B, Simmons P, Brouard N et al. Identification of stromal MSC

candidates from multiple adult mouse tissues. Int Soc Stem Cell Res 4th Annual

Meeting 2006, 206

158. Shi S, Gronthos S. Perivascular niche of postnatal mesenchymal stem cells in

human bone marrow and dental pulp. J Bone Miner Res 2003. 18:696-704

159. Sorrentino A, Ferracin M, Valtieri M et al. Isolation and characterization of

CD146+ multipotent mesenchymal stromal cells. Exp Hematol 2008. 36:1035-1046

160. Sacchetti B, Funari A, Michienzi S et al. Self-renewing osteoprogenitors in bone

marrow sinusoids can organize a hematopoietic microenvironment. Cell 2007.

131(2):324-336

161. Crisan M, Yap S, Casteilla L et al. A perivascular origin for mesenchymal stem

cells in multiple human organs. Cell Stem Cell 2008. 3:301–313

162. da Silva Meirelles L, Chagastelles P, Nardi N. Mesenchymal stem cells reside in

virtually all post-natal organs and tissues. J Cell Sci 2003. 119:2204-2213

163. Kolf C, Cho E, Tuan R. Biology of adult mesenchymal stem cells: regulation of

niche, self-renewal and differentiation. Arthritis Res and Therapy 2007. 9:204-214

164. Bianco P, Sacchetti B, Riminucci M. Osteoprogenitors and the hematopoietic

microenvironment Best Practice and Res Clin Hematol 2011 24:37-47

61 165. Díaz-Flores L, Gutiérrez R, Madrid J, et al. Pericytes. Morphofunction,

interactions and pathology in a quiescent and activated mesenchymal cell niche.

Histol Histopathol (2009) 24: 909-969

166. Bianco P, Riminucci M, Gronthos S, Gehron Robey P. Bone marrow stromal

cells: nature, biology, and potential applications. Stem Cells 2001 19:180-192

167. Westen, H. and D.F. Bainton, Association of alkaline-phosphatase-positive

reticulum cells in bone marrow with granulocytic precursors. J Exp Med, 1979.

150(4):919-37.

168. Bianco, P. and A. Boyde, Confocal images of marrow stromal (Westen-Bainton)

cells. Histochemistry, 1993. 100(2):93-9.

169. Bianco, P., et al., Marrow stromal (Western-Bainton) cells: identification,

morphometry, confocal imaging and changes in disease. Bone, 1993. 14(3):315-20.

170. Bianco, P. and G. Cossu, Uno, nessuno e centomila: searching for the identity of

mesodermal progenitors. Exp Cell Res, 1999. 251(2):257-63

171. Owen, M., Lineage of osteogenic cells and their relationship to the stromal

system. Bone Miner Res, 1985. 3: p. 1-25.

172. Owen, M. and A.J. Friedenstein, Stromal stem cells: marrow-derived osteogenic

precursors, in Cell and Molecular Biology of Vertebrate Hard Tissues. 1988, Wiley:

New York, NY. p. 42-60.

173. Fell, H.B., The osteogenic capacity in vitro of periosteum and endosteum isolated

from the limb skeleton of fowl embryos and young chiks. J Anat, 1932. 66: p. 157-80.

62 174. Nakahara, H., V.M. Goldberg, and A.I. Caplan, Culture-expanded human

periosteal-derived cells exhibit osteochondral potential in vivo. J Orthop Res, 1991.

9(4): p. 465-76.

175. O'Driscoll, S.W., Articular cartilage regeneration using periosteum. Clin Orthop,

1999. 367(Supple): p. S186-S203.

176. Patt HM, Maloney MA. Relationship of bone marrow cellularity and proliferative

activity: a local regulatory mechanism. Cell Tissue Kinet 1972. 5(4):303-309

177. Patt HM, Maloney MA. The bgJ/bgJ:W/WV bone marrow chimera: A model for

studying stem cell regulation. Blood Cells 1978. 4(1-2):27-35

178. Maloney MA, Lamela RA, Patt HM. The question of bone marrow stromal

fibroblast traffic. Ann NY Acad Sci 1985. 459:190-197

179. Jones E, English A, McGonagle D et al. Large scale extraction and

characterization of CD271+ multipotential stromal cells from trabecular bone in

health and osteoarthritis. Arth and Rheum 2010 62:1944-1954.

63 Chapter 2:

Introduction to Cell Separation Methods and Cell Surface Marker

Characterization of CTPs

Cell separation methods provide powerful tools for discriminating between diverse cell types and investigating cells for various surface markers. Within a mixture of cells, a specific cell type can be identified using previously established markers; furthermore, these specific cells can be isolated to produce a purified population. Alternatively, a single known cell type can be screened for a new potential marker. These tools provide methods for concentrating or isolating cell populations of interest to both researchers and clinicians.

Tools for cell isolation are especially useful when applied to cell suspensions containing many types of cells or when the cell type of interest is of low prevalence. As noted in

Chapter 1, the osteogenic cells resident in bone marrow, termed connective tissue progenitors, or CTPs, are the desired cell type for bone regeneration. However, CTPs among nucleated cells in marrow aspirates are very rare, on average 1 in 20,000 cells1,2, and this rarity highlights the need for separation technologies to select these cells from the heterogeneous mix of cell types present in a marrow aspirate.

2.1 Use of autologous bone marrow in bone defects

The use of autologous bone marrow as a tool for repair of bone defects was pioneered by

Connolly in 1989. Connolly injected aspirated bone marrow (BMA) without a scaffold

64 into bone defects created in a delayed union rabbit model. Defects injected with BMA showed greater calcification and healed bone rigidity than controls (which received no injection).3 Connolly et al. went on to show that concentrating marrow-derived cells using density separation4 might be used to improve clinical performance.3-6 Other small uncontrolled clinical studies corroborated the efficacy of bone marrow to heal bone defects.7-9

2.2 Density gradient separation methods for cell separation

Hernigou et al. used density separation for concentration of bone marrow derived cells for bone healing and were the first to correlate the number of CTPs injected to the bone healing outcome.10 They injected concentrated BMA, using a Cobe centrifuge, without a scaffold into the site of 60 human tibial nonunions. Union was achieved in 53/60 cases with an average time to union of 12 weeks. A positive correlation was found between the number of CTPs and volume of callus. It was noted that the 7 nonunion patients also had the lowest number of CTPs (<1000/cm3) and suggest a minimum of 1000 CTPs/cm3 was necessary to achieve union.10

Density separation using a centrifuge had been performed in other studies to heal these types of defects with success.3-6,10-13 Density separation uses the inherent differences in cell density to eliminate red blood cells (RBCs) and polymorphonuclear leukocytes that are of a higher density and pellet to the bottom, while recovering nucleated cells in a concentrated buffy coat layer on top of the pelleted RBCs. The advantage of density separation (DS) is that it can be used to rapidly process large numbers of cells and can be

65 used intraoperatively in the OR through modification of centrifuge devices designed for preparation of platelet rich plasma (PRP).10,12-13 Examples of these clinically available devices include: the Angel (Cobe), the SmartPReP2 APC+TM (Harvest Technologies), the AccelerateTM (Exactech), and the MarrowXpressTM (Thermogenesis).

DS can increase the cell and CTP concentration by elimination of the RBCs and harvest of the small volume of the buffy coat layer. DS is limited by its non-selectivity and it lacks the opportunity to increase CTP prevalence from that of native marrow. DS methods concentration nucleated marrow cells, including CTPs, but do not offer any method for selection of CTPs from the other nucleated cells in a BMA. DS methods often serve as an initial preprocessing step for marrow cells to remove the bulk of the RBCs before application of a selection method for CTPs.

2.3 Selective Retention methods for cell separation

In addition, “selective retention”, has also been described by Muschler et al. as a means of rapid intra-operative processing enabling both concentration and selection of marrow- derived CTPs.14-16 Selective Retention (SR) involves passing a cell suspension through a porous matrix and uses the natural affinity of CTPs for the matrix surface as a means to concentrate and select CTPs from the mixed population. Non-adherent cells are not retained and pass through the matrix. Unlike DS systems, SR results in a rapid and significant increase in both the concentration and prevalence of CTPs within the matrix.

Using SR methods, CTPs are preferentially retained compared to both RBCs and other nucleated cells that might compete with transplanted CTPs and compromise their survival

66 and function in the wound site. This process is directly analogous to an affinity column and in clinical application the biomaterial matrix in the column also becomes the implantable delivery system for the retained cells.15

Muschler et al. demonstrated that SR of marrow-derived cells using canine demineralized allograft cortical bone powder (DMBP) enhanced the performance of DBMP in the setting of canine spinal fusion.16 Union rate and union score was significantly greater than that achieved with DBMP alone or DBMP combined with unprocessed bone marrow and were comparable to the performance of autogenous cancellous bone.16 Other laboratories independently confirmed the efficacy of selective retention processing.17 One advantage of SR is minimal manipulation of the cells is required other than anticoagulation. SR processing also has the advantage of requiring relatively simple instrumentation and a minimum of reagents. However, even in retained “CTP enriched” populations, CTPs represent only a small fraction of the total number of retained cells. Using SR methods to load BMA onto DMBP has resulted in a cell retention efficiency of 23 ± 8% and a CTP retention efficiency of 61± 14%, giving a selection ratio of CTPs to other nucleated cells of 3.0 ± 1.5. However, due to the low frequency of CTPs, the number of nucleated cells implanted was 269± 41 million cells, while the number of CTPs implanted was 39,400 ±

24,500.16 Further improvement in SR processing will require more selective surfaces or other means of eliminating competing cells.

In May of 2006, Depuy Spine launched the first clinical product that was designed to enable surgeons to utilize selective retention processing in the operating room. Over two

67 years approximately 20,000 patients received bone grafts processed with the Depuy

Cellect™ Device 18,19 until the FDA reclassified this device as Class III in 2009. The overall success rate that has been reported from several small series has been 85-95%.18

2.4 Fluorescence activated cell sorting (FACS) for cell separation

Fluorescence activated cell sorting (FACS) provides greater versatility for marrow cell selection, based on cell morphology or the presence of one or more surface antigens.

FACS uses differences in cell size and granularity to separate unlabeled cells, as lymphocytes, monocytes, and neutrophils form distinct discrete clusters. Cells can also be tagged with fluorescent antibodies for any cell surface marker and separated based on the intensity of the fluorophore. Multiple fluorescent labels can be used to fractionate different cell populations, and sterile sorting methodology is employed to enable cell culture after separation. FACS is a powerful tool for cell separation because of its diagnostic, prognostic and predictive significance. 20-22 It is especially useful for rare cell analysis. For example, flow cytometry is used to analyze peripheral blood samples for rare circulating tumor cells. Cells can be screened for markers that are predictive of clinical prognosis.20-22

However, clinical use of FACS is limited by cost, issues of sterility, and throughput limitations. It remains primarily an analytical tool, as a method for assessment of other clinical scale processing methods or as a diagnostic tool, than as a means of clinical cell processing itself.

68 2.5 Magnetic separation technologies for cell separation

Magnetic separation (MS) uses the technique of coupling magnetic beads to antibodies for cell surface markers. Cells tagged in this way and placed into an external magnetic field will move toward the magnet wall and be retained by the magnetic field, while unlabeled cells in the mixture will remain in suspension. A cell may be labeled with no beads to multiple beads, in proportion to the density of the antigen, and will be accelerated in the magnetic field in proportion to the number of beads bound to its surface. This “negative” unlabeled fraction can be removed and a purified population remains.

Magnetic separation addresses some of the limitations of flow cytometry, namely, higher throughput and the elimination of the high shear forces present in the cytometer.

Commercially available magnetic separation systems can manage 100 million to 1 billion cells in a small permanent magnet. However, MS methodology are generally limited to binary separation (i.e. positive or negative cell fractions only), and can be insensitive to the degree of cell magnetization. The same animal-derived antibody biocompatibility issues that affect the clinical use of flow cytometry also affect the utility of MS processing. Both systems can select a wide range of cell types using customizable beads.

Magnetic separation systems offer the advantages of ease-of-use, convenient and simple instrumentation, and increased viability of “fragile” cells.

2.6 Magnetic Theory

Generally, when a material is placed in an external magnetic field of strength H, it responds on an atomic level. This response is referred to as the magnetic flux density B

69 (also called the magnetic induction or magnetic field intensity), which is an inherent property of the material.23 The permeability of the material  relates the material response

B with the external field H. In a vacuum,

B   o H

However, in any other situation,  may vary with the external field, and the relationship between B and H may not be a simple linear function. Magnetization M is defined as the magnetic moment m divided by the volume of the magnetic material V. Magnetization represents the cumulative effect of all the individual magnetic dipole moments in a material in a given volume.

m M  V

Now the magnetization relates the magnetic properties of the material to the magnetic induction and the applied field, and the complete equation representing the response of a material to the applied external field is:

B  o (H  M )

The above equation is used magnetic induction in any medium that is not a vacuum. The magnetic susceptibility  of a material is the degree of magnetization of a material in response to an external magnetic field and is defined by

M   H

The response of a material to an applied magnetic field is dependent on its atomic magnetic moments, which themselves depend on electron orbits and the chemistry and

70 balance of electrons in their shells.23 These factors determine whether a material is ferromagnetic, ferrimagnetic, diamagnetic, or paramagnetic.

2.7 Types of Magnetism

Ferromagnetic materials include all permanent magnets, for example, iron and nickel. On an atomic level these materials have multiple electrons with unpaired spins. It has been postulated that it is energetically favorable (lower energy) for the unpaired electrons‟ magnetic moments to align and become parallel to each other even in the absence of an applied magnetic field. In addition, adjacent atom‟s magnetic moments will align parallel to each other and form whole magnetic domains within the material that are aligned together and form large magnetic fields. However, the material may be composed of thousands of domains that all are aligned differently with respect to each other, so that when averaged, the net magnetization for the material is zero. When placed in an applied magnetic field, the domains that were already aligned with the direction of the applied field will enlarge, and other domains will reorient themselves to align with the applied field to reach a lower energy configuration. Both of these actions will work together to produce a large net magnetization. Once a ferromagnetic material is magnetized by a magnetic field it remains magnetized after field is removed, showing magnetic “memory”.

Ferrimagnetism works under the same principles of ferromagnetism except that some of it‟s domains in the applied field are oriented parallel and some are anti-parallel. They are uneven in magnitude, so there is an overall positive net magnetization, which is smaller

71 than that of a ferromagnetic material. This imbalance of domains is due to presence of Fe ions in different oxidation states.

Paramagnetic materials‟ molecules contain stable magnetic dipole moments attributable to the magnetic moments of unpaired electrons, but these materials do not form domains like ferromagnetic materials. Due to the randomizing effects of thermal motion, these moments are normally randomly oriented so that the material has no net magnetization.

In an applied magnetic field, the net magnetic moments align along the direction of the magnetic field and a weak but positive net magnetization is seen. These materials have no retained magnetism when the field is removed, as randomizing temperature effects once again dominate.24

Superparamagnetic materials can be generated from ferro- and ferri-magnetic materials.

When manufactured into particles below a critical diameter (usually 1-30 nm, material specific), these particles will behave differently than the bulk material. The magnetic moments present in the particle cannot align because they are constantly being disrupted and realigned by random thermal energy. The particles themselves are now nonmagnetic outside of an applied magnetic field. However, in an external field those magnetic domains will overcome the randomizing effects of thermal energy and align in the field.

When removed from the field they exhibit no “memory” and the net overall moment returns to zero. Now these tiny ferromagnets are behaving like paramagnets. However, these particles will have very large magnetic susceptibilities – higher than that of paramagnets – and are therefore termed superparamagnetic particles.23

72 2.8 Clinical Application of Magnetic Separation Systems

Magnetic separation strategies are currently used for progenitor cell enrichment for bone marrow transplantation in cancer patients and detection of cancer cells in circulating blood. Transplantation of hematopoietic stem cells (HSC) from bone marrow is often used in the treatment of hematologic and lymphoid cancers. HSCs, mobilized from bone marrow and circulating in peripheral blood, are selected based on their expression of the cell surface marker CD34.25,26 For example, Lang et al. mobilized HSC using granulocyte colony-stimulating factor (G-CSF) and harvested CD34+ cells from peripheral blood using CliniMACS.26 The CD34+ cell suspension was infused into pediatric leukemic patients. The median purity was 98.5% in CD34+ fraction. This method depleted t-cells, responsible for the development of graft vs host disease (GVHD), from the transplanted cell fraction. The patients did not receive immunosuppressive therapy after these infusions. When compared to a historical control group who received unmanipulated bone marrow, the patients who received the CD34+ MS-processed cells had a significantly lower incidence of GVHD. However, there was no difference was in the

26 probability of relapse or survival.

Clinically approved magnetic separation systems that have been used to isolate hematopoetic stem cells mobilized to peripheral blood include the CliniMACS system

(Miltenyi Biotec) and the Isolex system (Nexell Therapeutics Inc). The CliniMACS system is CE-marked (Conformité Européenne) for clinical use in Europe. In the US,

CliniMACS products are available only under an approved Investigational New Drug

(IND) application or Investigational Device Exemption (IDE) for clinical research use. It is noted that CliniMACS MicroBeads are for research use only and not for use in humans.

73 The Isolex system was approved through a PMA in 1997, and uses paramagnetic microbeads coupled to a CD34 antibody to isolate CD34+ cells.27

2.9 Advantages and disadvantages of commercially available magnetic separation systems

Invitrogen Inc. markets the Dynal Magnetic Particle Concentrator (MPC) which was the first system with magnetic particles used for clinical cell separation.28 The MPC is paired with Dynabeads magnetic particles, which are micrometer size beads (~4.5 micrometer in diameter). The advantage of this system is the pairing of the strong magnetic bead with a large magnetic moment, which enable the use of a weaker magnet with an open gradient for separation. In this system, cells are labeled with Dynabeads and placed in the magnet.

Cells labeled with Dynabeads will be retained in the magnet against the wall and unlabeled cells can be decanted. However, this system has several disadvantages including slow reaction kinetics, the tendency for the large beads to precipitate out of solution, the potential for nonspecific cell entrapment due to aggregation of beads, and the potential for cellular damage (the cell experiences large shear forces as the bead moves toward magnet wall due to large magnetic moment of the bead).25 As a result, the

MPC system is often used for depletion of undesired cells, also called negative selection, which eliminates the issue of cellular damage to the positively labeled cells.

Miltenyi Biotec markets the popular MACS systems. These magnetic beads used for

MACS separation are superparamagnetic nanometer size beads (100 nm) with a biodegradable dextran matrix and fast reaction kinetics. These beads are in a colloidal suspension and do not precipitate like the Dynal beads, but their magnetic moment is

74 much weaker.25 The MACS beads must be used with a strong magnet in a specialized column. This “affinity” style column contains soft steel alloy inserts that produce a locally high gradient magnetic field. A high surface area porous column of ferromagnetic wire mesh or sub-millimeter beads is used.29 When placed in a magnetic field, a local fringe field gradient is produced near the surface of the matrix (extending a few cell diameters into solution).29 Cells are labeled and flow through the column. Positive cells are magnetically retained against the steel “beams” in the column while negative cells flow through. Positive cells are subsequently eluted from column after removal from magnet. The nature of the affinity style column has several disadvantages. The column can overload and clog, and the surface area for cell retention is fixed and can saturate when trying to separate large numbers of positive cells. The column cost is much higher due to its specialized nature. Complete removal of all retained cells after the separation can be a challenge, especially for “sticky” cells that adhere to many surfaces.

Stem Cell Technologies produces the EasySep System, consisting of an open gradient quadrupole magnet and nanometer sized beads (~160 nm). Like the MACS system, these are superparamagnetic colloidal dextran-coated magnetite bead suspensions. Cells are labeled with the EasySep tetrameric antibody complex (a double sided antibody specific for the ligand of interest on one side and dextran on the other). Similar to the Dynal system, labeled cells are placed in the magnet and unlabeled cells are poured off while positive cells are retained in tube in magnet. The compact size of the magnet and small volume required for labeling and separation make the EasySep magnetic system useful as a tool for screening cell surface markers for CTP selection. The EasySep system was

75 selected and used in the small scale proof-of-concept experiments that evaluated cell- associated hyaluronan as a marker to enrich CTPs.

The magnetic beads used to tag the cells in the EasySep system are dextran-coated magnetite. Magnetite (Fe3O4, or FeO*Fe2O3) is a naturally occurring mineral that is ferrimagnetic. However, it can be processed into small particles that behave paramagnetically but retain their large susceptibility, and thus the beads used in the

EasySep system are defined as superparamagnetic particles. A diagram of a cell tagged with these magnetic particles is presented in figure 2.1. Fresh human bone marrow aspirates were stained sequentially with a biotinylated hyaluronic acid binding protein, followed by the EasySep tetrameric antibody complex. The dextran-coated magnetic beads are then added and the cells are fully labeled and ready for placement in the magnet.

Figure 2.1: Example of staining sequence with EasySep reagents using hyaluronan as a marker for selection.

2.10 Geometry of the Magnet

The EasySep magnet is a small permanent magnet that accommodates a 5mL polystyrene falcon tube (diameter = 11mm) containing the magnetically labeled cell suspension. The

76 EasySep magnet utilizes a quadrupole geometry, indicating that there are 4 permanent magnets oriented at 0o, 90o, 180o, and 270o. A quadrupole magnet offers compact design while generating high magnetic fields and high gradients. An ideal quadrupole field produces only a centrifugal force field, as B is dependent on radial position r only.25

2.11 Cell surface markers used to select CTPs

Given the low frequency of CTPs in bone marrow, there is a need for clinically relevant selection strategies to optimize the impact of these cells in a graft site. CTPs comprise a heterogeneous population of cells that encompass stem and more committed progenitor cells, and the cell surface marker expression may also be heterogeneous on this population. A definitive surface marker profile has yet to be established for the CTP population in native tissues, including bone marrow, making identification of these cells in vivo or using rapid isolation a continuing challenge.

Due to the relative infrequency of these cells, many investigators have focused on phenotyping the culture expanded progeny of CTPs, termed mesenchymal stem cells

(MSCs, see Chapter 1). However, culture conditions can alter cell surface marker expression and select for a more homogeneous phenotype and subset of CTPs. Some markers are only expressed on cultured MSCs, for example, CD109, CD166, and

CD318.30,31 Other markers are expressed on primary, or “native” MSCs (which are also referred to as multipotent mesenchymal stromal cells , mesenchymal progenitor cells, marrow stromal progenitor cells, mesenchyal stem cells, or stromal cell precursors) but are rapidly downregulated on the same cells during the culture period.31 Researchers often report differential expression of markers on native and cultured MSCs. Markers that

77 are positive on native MSCs, including CD34,32 CD45,33,34 CD133,35 and CD271,33 have been found by other researchers to be negative on cultured MCSs (CD34,36 CD45,36

CD133,35 CD27137).38 It is unknown if these markers are initially present but downregulated during culture, or if the culture conditions themselves have selected for a subpopulation with different expression than the founding CTPs. There is also conflicting data on some markers, for example, CD45 and CD34, which some researchers have found to be positive on native MSCs and others have not.32-34,39,40,42

Given the confusion around the various names ascribed of these cells and the varying definitions of an MSC postulated by different researchers, the International Society for

Cellular Therapy recently proposed a formal nomenclature of multipotent mesenchymal stromal cells (MSCs), and three-part criteria for defining an “MSC”, which includes the typical assay for CFU by plastic adherence, tri-lineage differentiation, and expression of a panel of surface markers.40,41 They proposed that MSC express CD105, CD73, and CD90.

Since each of these markers are also present on other cells in bone marrow, ISCT includes testing that MSC are negative for HLA-DR, CD45, CD34, CD14 or CD11b, and

CD79a or CD19. 40,41

Three of the first antibodies to react specifically with surface proteins on MSCs were

SH2, SH3 and SH4.43 It was later discovered that the antibody SH2 recognizes CD10544 and that SH3 and SH4 recognize CD73.45 CD90 has also been found on MSCs.34 The markers that MCSs are expected to be negative for are largely markers that are found on hematopoietic cells, which have a much higher prevalence in bone marrow aspirates, and that may “contaminate” MSC cultures during the plastic adherence assay. CD45 is a pan-

78 leukocytic marker that detects all mature hematopoietic cells. CD34 is found on primitive hematopoietic progenitors and endothelial cells. CD14 and CD11b are expressed on monocytes and macrophages, which are the most common hematopoietic cells found in

MSC culture. CD79a and CD19 are markers of B cells, and HLA-DR molecules are not expressed on MSC (unless the MSCs are previously stimulated to express it). 40,41

Characterizing the cell surface protein expression of MSCs has been an active area of research for 20 years now. Unfortunately, identification of CTPs and MSCs is retrospective, based on their adherence to tissue culture plastic and ability to proliferate and form colonies. Characterization of these cells has thus been performed largely on cultured cells. Since cultured (and especially passaged) MSCs are often a homogeneous population due largely to their culture conditions, the characterization of these cells is not reflective of the heterogenous native CTP population.

Methods for selection of CTPs that are appropriate for rapid introperative processing, the focus of the Muschler lab, must be prospective. Any marker selected for CTPs must be present on at least a subset of freshly aspirated BM-derived CTPs, and not present (or present on a minimal fraction) on the more numerous non-osteogenic cells in a BMA.

Therefore, the cell surface markers presented on the following pages have been identified as potential prospective markers for the native human CTP population that have been isolated from bone marrow. Table 2.1 summarizes these findings, and includes the level of CFU enrichment in the positive fraction over the unselected marrow. Details on some of the more commonly used markers are summarized below.

79 Author Marker Molecule Quirici CD271 low affinity nerve growth factor receptor Martinez GD2 neural ganglioside Gang SSEA-4 stage specific embryonic antigen 4 Campioni CD105 endoglin, receptor for transforming growth factor beta III Majumdar CD105 endoglin, receptor for transforming growth factor beta III Sacchetti CD146/CD45 melanoma-associated cell adhesion molecule Deschaseaux CD49a a1-integrin subunit from very late antigen 1 (VLA-1) integrin Gindraux CD49a a1-integrin subunit from very late antigen 1 (VLA-1) integrin Gindraux CD49a a1-integrin subunit from very late antigen 1 (VLA-1) integrin Gindraux CD49a then CD133 a1-integrin subunit from very late antigen 1 (VLA-1) integrin Simmons STRO-1 unknown Simmons STRO-1/GLyA unknown, glycophorin A Gronthos STRO-1 unknown Gronthos STRO-1 then VCAM-1 unknown, vascular cell adhesion molecule-1 Shi STRO-1 then CD146 unknown Delorme CD49b a2-integrin subunit from very late antigen 1 (VLA-1) integrin Delorme CD105 endoglin, receptor for transforming growth factor beta III Delorme CD90 Thy-1 (THYmocyte differentiation antigen 1) Delorme CD73 ecto-5'-nucleotidase Delorme CD130 subunit of type 1 cytokine receptor (IL-6) Delorme CD146 melanoma-associated cell adhesion molecule Delorme CD200 cell surface glycoprotein Delorme AVB5 alpha4beta5

Fraction Enriched Preprocessing method Separation method Percent of cells + in BM CFU enrichment level Reference CD271+ Ficoll MACS 2.3± 0.8% 61-fold ↑ 33 not reported not reported MACS not reported not reported 46 SSEA-4+ density separation FACS 2-4% not reported 47 CD105+ Lympholyte MACS 1.1% 2.7 fold ↑ 48 CD105+ Ficoll MACS 1.8% 9-fold ↑ 49 CD45-/CD146+ none listed MACS then FACS 0.11% 830-fold ↑ 39 CD49a+ Ficoll MACS 3.6% 20-fold ↑ 34 CD49a+ Ficoll Dynal 3.6 21-fold ↑ 35 CD49a+ Ficoll MACS 0.4 143-fold ↑ 35 CD49a+/CD133+ Ficoll MACS then FACS 0.3 100-fold ↑ 35 STRO-1+ Ficoll FACS ~10% 20-fold ↑ 50 STRO-1+/GLyA- Ficoll FACS 5% of STRO-1+ cells 100-fold ↑ 50 STRO-1BRT BMMNCs MACS then FACS 6.50% 950-fold ↑ 51 STRO-1BRTVCAM-1+ BMMNCs MACS then FACS 1.4% of STRO-1+ cells CFU: 1 in every 3 cells 51 STRO-1BRTCD146+ purchased BMMNCs MACS then FACS not reported 2000-fold ↑ 52 CD49b+ "BM MNCs", undefined FACS 0.7 23-fold ↑ 53 CD105+ "BM MNCs", undefined FACS 1.9 50-fold ↑ 53 CD90+ "BM MNCs", undefined FACS 0.28 60-fold ↑ 53 CD73+ "BM MNCs", undefined FACS 0.16 100-fold ↑ 53 CD130+ "BM MNCs", undefined FACS 0.17 256-fold ↑ 53 CD146+ "BM MNCs", undefined FACS 0.16 278-fold ↑ 53 CD200+ "BM MNCs", undefined FACS 0.15 333-fold ↑ 53 AVB5+ "BM MNCs", undefined FACS 0.014 1750-fold ↑ 53

Table 2.1: Cell surface markers used for the prospective isolation of native bone marrow- derived progenitor cells.

80 2.11.1 Selection based on STRO-1

One of the first antibodies to be used for prospective identification of MSCs, Simmons and Torok-Straub first reported the use of STRO-1 to select MSCs in 1991.50 Despite the fact that STRO-1 has been researched and used for cell selection for 20 years, the antigen to this antibody is still unknown. STRO-1+ cells include a small subpopulation of cell that are MSCs, as well as nucleated erythroid cells (about 90% of STRO-1+ cells) and B- lymphocytes.51

Simmons et al. used the STRO-1 antibody to label bone marrow mononuclear cells

(BMMNCs) and separated cells using FACS. Approximately 10% of BMMNCs were

STRO-1+. All colonies formed were in STRO-1+ fraction, and the level of enrichment of

MSCs over the BMMNCs was 20-fold.50

Because over 90% of STRO-1+ cells are non-MSCs, separation based on STRO-1 is often paired with another cell surface marker, either for further positive selection of

MSCs, or for depletion of non-MSCs from the STRO-1+ population. Simmons evaluated

STRO-1+ cells for glycophorin A, a marker for cells of the erythroid lineage, as an additional marker to deplete nucleated erythrocytes. About 5% of the STRO-1+ cells were gycophorin A-, and the level of enrichment of MSCs based on STRO-1+/GlyA- separation was 100-fold.50

Gronthos et al. paired STRO-1 isolation using MACS with subsequent FACS to select

STRO-1BRT/VCAM-1+ cells. Only 1.4% of STRO-1+ cells also expressed VCAM-1.

81 One in three STRO-1BRT/VCAM-1+ cells formed a colony in vitro, indicating an extremely high level of MSC enrichment.51 VCAM-1 is vascular cell adhesion protein 1, also known as CD106, and functions as a cell adhesion molecule. VCAM-1 is most commonly found on endothelial cells, where it regulates leukocyte migration across blood vessel walls and is used for attachment during angiogenesis.

Shi et al. used MACS selection on BMMNCs for STRO-1 followed by FACS for

CD146+ cells. Isolation of STRO-1BRT/CD146+ cells resulted in a 2000-fold enrichment of MSCs over BMMNCs. (1 CTP per 5 STRO-1BRT/CD146+ cells).52

2.11.2 Selection based on CD105:

CD105 is one of the markers recommended by ISCT to be included on the panel of markers for positive selection of MSCs40,41 and was one of the first markers identified for

MSCs by the antibody SH-2.43 CD105, also called endoglin, is a receptor for transforming growth factor beta III. It is known to be expressed by MSCS as well as endothelial and stromal cells in BM.48,49,54

Campioni et al. isolated CD105+ cells from BMMNCS using MACS. They found 1.1% of cells partitioned to CD105+ fraction with a purity of 66.5%. A 2.7-fold increase over

BMMNCs was found in CFU-F.48 Majumdar et al. also used MACS for CD105 separation and found using a directly conjugated antibody that 1.8% of cells partitioned to the CD105+ fraction. This resulted in a 9-fold increase in CTPs over BMMNCs.49

2.11.3 Separation based on CD49a

82 CD49a has been studied as a possible marker for native MSCs.34,35,55 The CD49a molecule is the a1-integrin subunit from very late antigen 1 (VLA-1) integrin. This adhesion molecule is specific to the vascular smooth muscle lineage and acts as a receptor for both collagen and laminin.34 Deschaseaux et al. performed magnetic separation on BMMNCs based on CD49a expression. The percentage of BMMNCs retained in the CD49a+ fraction was 3.6%. All CFU-F were recovered in the CD49a+ fraction after 10-14 days of culture. The level of enrichment in MSC in the CD49a+ fraction over the BMMNCs was 20-fold, and the recovery using the MACS column was

84%.34

CD49a (MACS) separation has been combined with CD133 (FACS) to isolate

CD49+/CD133+ cells, about 0.3% of the starting BMMNCs. This resulted in a 100–fold enrichment in MSCs over BMMNCs.35 CD133 is also found on hematopoietic stem cells.

2.11.4 Separation based on CD73

Antibodies SH-3 and SH-4 (as well as SH-2) were first used to isolate MSCs in 1992.43

They were found to react against cultured and native MSCs but not hematopoietic cells in marrow. It was later discovered that SH-3 and SH-4 bind CD73 on MSCs.45 CD73 is expressed on multiple cell types including endothelial cells and lymphoid cells where is plays a role in B-cell activation.45 CD73 is another marker recommended by ISCT to identify MSCs.40,41

2.11.5 Selection based on CD271

83 Cattoretti et al examined bone marrow biopsies for expression of CD271 (also known as low-affinity nerve growth factor receptor). They found CD271 labels most of the reticulin+ and collagen III+ stromal cells in BM, and did not label endothelial or hematopoietic cells.56 The function of CD271 in BM is unknown.

CD271 has gained popularity in recent years as a prospecitive marker for MCSs. Quirici et al performed magnetic separation using the MACS system based on CD271 on bone marrow mononuclear cells (BMMNCs) separated by Ficoll-PaqueTM. In the starting sample of BMMNCs, 2.3± 0.8% of cells were CD271+. After separation the purity as assessed by flow cytometry was 90.5± 3.5%. The recovery of CD271+ cells in the positive fraction was only 10% of the total number of cells that were CD271+. However,

CFU-F enrichment in CD271+ fraction was 61 fold compared to BMMNC, indicating that the CD271+ cells that were isolated contained a higher prevalence of progenitors than unprocessed BMMNCs.33

2.11.6 Separation based on CD146

CD146, or melanoma cell adhesion molecule, is a pericyte marker that has recently been identified as a marker for MSCs.39,52,53,57 It is also expressed on endothelial cells and t- cells.

Sacchetti et al. separated CD45-/CD146+ and CD45-/CD146- cells from bone marrow aspirates using MACS to isolate CD45- cells and FACS to partition the CD146 positive and negative fraction within the CD45- cell population. All colony forming cells were found in the CD45-/CD146+ fraction, which was an 830-fold enrichment over unselected

84 bone marrow. The CD45-/CD146+ cells also expressed alkaline phosphatase and

CD105.39 Crisan et al. performed fluorescence activated cell sorting on cells dissociated from multiple tissues, including bone marrow. Cells were first gated for CD56- to remove contaminating cells with myogenic potential, then CD45- to eliminate all hematopoietic cells, and finally selected for CD146+ CD34- cells. These CD56- CD45- CD146+ CD34- cells differentiated along the chondrogenic, adipogenic, and osteogenic lineages in vitro, and formed bony nodules in vivo when implanted into an immunocompromised mouse model.57

Delorme et al. separated cells using FACS based on multiple surface markers on the same patients, providing a comparison of CFU enrichment based on CD105, CD49b, CD90,

CD73, CD130, CD146, CD200, and alphaV/beta5 markers. The percentages of BM cells recovered from the total MNCs was 1.9% for CD105, and <1% from all other markers tested. After FACS sorting, cells were plated into CFU assay and the level of enrichment in CFU in the positive fractions over the BMMNCs was calculated. “Low” CFU enrichment was found based on CD49b, CD105, and CD90 separations (range:

23- to 60 -fold), “high” enrichment was found based on CD73, CD130, CD146, and CD200 (range: 100 - to 333 -fold), and “very high” enrichment (1750) was found with alphaV/beta 5 FACS separation.

85 2.12 References

1. Muschler, G.F., Boehm, C. & Easley, K. Aspiration to obtain osteoblast progenitor

cells from human bone marrow: the influence of aspiration volume. Journal of Bone

and Joint Surgery (American) 79, 1699-709 (1997).

2. Majors, A.K., Boehm, C.A., Nitto, H., Midura, R.J. & Muschler, G.F.

Characterization of human bone marrow stromal cells with respect to osteoblastic

differentiation. Journal of Orthopaedic Research 15, 546-57 (1997).

3. Connolly, J., et al., Development of an osteogenic bone-marrow preparation. J

Bone Joint Surg Am, 1989. 71(5): p. 684-91.

4. Connolly, J.F., et al., Autologous marrow injection as a substitute for operative

grafting of tibial nonunions. Clin Orthop, 1991. 266: p. 259-70.

5. Tiedeman, J.J., et al., Treatment of nonunion by percutaneous injection of bone

marrow and demineralized bone matrix. An experimental study in dogs. Clin

Orthop, 1991. 268: p. 294-302.

6. Connolly, J.F., Injectable bone marrow preparations to stimulate osteogenic repair.

Clin Orthop, 1995. 313: p. 8-18.

7. Garg, N.K. and S. Gaur, Percutaneous autogenous bone-marrow grafting in

congenital tibial pseudarthrosis. J Bone Joint Surg Br, 1995. 77(5): p. 830-1.

8. Garg, N.K., S. Gaur, and S. Sharma, Percutaneous autogenous bone marrow

grafting in 20 cases of ununited fracture. Acta Orthop Scand, 1993. 64(6): p. 671-2.

9. Healey, J.H., et al., Percutaneous bone marrow grafting of delayed union and

nonunion in cancer patients. Clin Orthop, 1990. 256: p. 280-5.

86 10. Hernigou, P., et al., Percutaneous autologous bone-marrow grafting for nonunions.

Influence of the number and concentration of progenitor cells. J Bone Joint Surg

Am, 2005. 87(7): p. 1430-7.

11. Hernigou, P., et al., Percutaneous implantation of autologous bone marrow

osteoprogenitor cells as treatment of bone avascular necrosis related to sickle cell

disease. Open Orthop J, 2008. 2: p. 62-5.

12. Hernigou, P., et al., The use of percutaneous autologous bone marrow

transplantation in nonunion and avascular necrosis of bone. J Bone Joint Surg Br,

2005. 87(7): p. 896-902.

13. Hernigou, P. and F. Beaujean, Treatment of osteonecrosis with autologous bone

marrow grafting. Clin Orthop, 2002. 405: p. 14-23.

14. Muschler, G.F., et al., Spine Fusion Using Cell Matrix Composites Enriched in

Bone Marrow-Derived Cells. Clin Orthop Rel Res, 2003. 407: p. 102-118.

15. Muschler, G.F., Method for Preparing a Composite Bone Graft. US patent No.

5,824,084, October 20. 1998.

16. Muschler, G.F., et al., Selective Retention of Bone Marrow-Derived Cells to

Enhance Spinal Fusion. Clin Orthop, 2005(432): p. 242-251.

17. Brodke, D., et al., Bone grafts prepared with selective cell retention technology heal

canine segmental defects as effectively as autograft. J Orthop Res, 2006. 24(5): p.

857-66.

18. Youssef, J., et al. Selective Cell Retention Technology in Spinal Fusion. in 18th

Annual Meeting of North American Spine Society. 2003. San Diego, CA.

87 19. Laurencin, C.T., Bone grafts and bone substitutes, in Orthopaedic Network News.

1999. p. 8-9.

20. Yasuda I, Goto N, Tsurumi H, Nakashima M, Doi S, Iwashita T, Kanemura N,

Kasahara S, Adachi S, Hara T, Shimizu M, Takami T, Moriwaki H. Endoscopic

Ultrasound-Guided Fine Needle Aspiration Biopsy for Diagnosis of

Lymphoproliferative Disorders: Feasibility of Immunohistological, Flow

Cytometric, and Cytogenetic Assessments. Am J Gasterenterol 2011 (Epub)

21. Yuan CM, Stetler-Stevenson M. Role of flow cytometry of peripheral blood and

bone marrow aspirates in early myeloma. Semin Hematol. 2011 48(1):32-8.

22. Peters JM, Ansari MQ. Multiparameter flow cytometry in the diagnosis and

management of acute leukemia. Arch Pathol Lab Med. 2011 135(1):44-54.

23. Hatch, G.P. & Stelter R.E. Magnetic design considerations for devices and particles

used for high-gradient magnetic separation (HGMS) systems Journal of Magnetism

and Magnetic Materials 225, 262-276 (2001)

24. Fishbane P., Gasiorowics S. & Thornton S. Physics for Scientists and Engineers. 2nd

Edition. Prentice Hall 1993 863-879

25. Zborowski M. & Chalmers J.J., Magnetic Cell Separation: A volume in laboratory

techniques in biochemistry and molecular biology. Elsevier 2007

26. Lang P, Handgretinger R, Niethammer D, et al. Transplantation of highly purified

CD34+ progenitor cells from unrelated donors in pediatric leukemia. Blood.

2003;101:1630-1636.

88 27. Isolex: Summary of safety and effectiveness document. Downloaded from FDA at:

http://www.fda.gov/downloads/biologicsbloodvaccines/bloodbloodproducts/approv

edproducts/premarketapprovalspmas/ucm091901.pdf

28. Kemshead J, Ugelstad J Magnetic separation techniques: their application to

medicine. Molecular and Cellular Biochem 1985. 67:11-18

29. Kantor AB, Gibbons I, Miltenyi S, Schmitz J. Magnetic cell sorting with colloidal

superparamagnetic particles. In: Recktenwald D, Radbruch A, eds. Cell Separation

Methods and Applications. New York: Marcel Dekker, Inc.; 1998:153-173.

30. Battula VL, Treml S, Bareiss PM, et al. Isolation of functionally distinct

mesenchymal stem cell subsets using antibodies against CD56, CD271, and

mesenchymal stem cell antigen-1. Haematologica 2009;94:173–84.

31. Vogel W, Grunebach F, Messam CA, et al. Heterogeneity among human bone

marrow-derived mesenchymal stem cells and neural progenitor cells.

Haematologica 2003;88:126–33.

32. Simmons PJ, Torok-Storb B. CD34 expression by stromal precursors in normal

human adult bone marrow. Blood.1991; 78: 2848–53

33. Quirici N, Soligo D, Bossolasco P, Servida F, Lumini C, Lambertenghi Delitiers G.

Isolation of bone marrow mesechymal stem cells by anti-nerve growth factor

receptor antibodies. Exp Hematol 2002 30:783-791

34. Deschaseaux F, Gindraux F, Saadi R, et al. Direct selection of human bone marrow

mesenchymal stem cells using an anti-CD49a antibody reveals their CD45med,low

phenotype. Br J Haematol. 2003; 122: 506–17.

89 35. Gindraux F, Selmani Z, Obert L, et al. Human and rodent bone marrow

mesenchymal stem cells that express primitive stem cell markers can be directly

enriched by using the CD49a molecule. Cell Tissue Res. 2007; 327: 471–83.

36. Pittenger MF, Mackay AM, Beck SC, et al. Multilineage potential of adult human

mesenchymal stem cells. Science. 1999; 284: 143–7.

37. Jones E, McGonagle D. Human bone marrow mesenchymal stem cells in vivo.

Rheumatology. 2008; 47: 126–31.

38. Deschaseaux F, Pontikoglou C, Sensébé L Bone regeneration: the stem/progenitor

cells point of view J. Cell. Mol. Med. 2010 14:103-115

39. Sacchetti B, Funari A, Michienzi S et al. Self-renewing osteoprogenitors in bone

marrow sinusoids can organize a hematopoietic microenvironment. Cell 2007.

131(2):324-336

40. Dominici M, Le Blanc K, Mueller I, et al. Minimal criteria for defining multipotent

mesenchymal stromal cells. The International Society for Cellular Therapy position

statement Cytotherapy (2006) Vol. 8, No. 4, 315-317

41. Horwitz E, Le BK, Dominici M et al . Clarification of the nomenclature for MSC:

the International Society for Cellular Therapy position statement. Cytotherapy

2005;7:393-5.

42. Kaiser S, Hackanson B, Follo M, et al. BM cells giving rise to MSC in culture have

a heterogeneous CD34 and CD45 phenotype. Cytotherapy 2007;9:439–50.

43. Haynesworth, S. E., Baber, M. A., and Caplan, A. I. Cell surface antigens on

human marrow-derived mesenchymal cells are detected by monoclonal antibodies.

Bone 1992 13, 69–80.

90 44. Barry, F. P., Boynton, R. E., Haynesworth, S., Murphy, J. M., and Zaia, J. (1999)

The monoclonal antibody SH-2, raised against human mesenchymal stem cells,

recognizes an epitope on endoglin (CD105). Biochem. Biophys. Res. Commun. 265,

134–139

45. Barry F, Boynton R, Murphy M, et al. The SH-3 and SH-4 antibodies recognize

distinct epitopes on CD73 from human mesenchymal stem cells. Biochem Biophys

Res Commun. 2001; 289: 519–24.

46. Martinez C, Hofmann TJ, Marino R, et al. Human bone marrow mesenchymal

stromal cells express the neural ganglioside GD2: a novel surface marker for the

identification of MSCs. Blood. 2007; 109: 4245–8

47. Gang EJ, Bosnakovski D, Figueiredo CA, et al. SSEA-4 identifies mesenchymal

stem cells from bone marrow. Blood.2007; 109: 1743–51.

48. Campioni D, Lanza F, Moretti S et al. Functional and immunophenotypic

characteristics of isolated CD105+ and fibroblast+ stromal cells from AML:

implications for their plasticity and endothelial lineage.

49. Majumdar M, Banks V, Peluso D and Morris E. Isolation, characterization and

chondrogenic potential of human bone marrow-derived multipotential stromal cells.

J of cell physiology 2000 185:98-106

50. Simmons PJ, Torok-Storb B. Identification of stromal cell precursors in human

bone marrow by a novel monoclonal antibody, STRO-1. Blood. 1991; 78: 55–62.

51. Gronthos S, Graves S and Simmons P. Isolation, purification, and in virto

manipulation of human bone marrow stromal precursor cells. 1998 In: Beresford J,

91 Owens M (eds.) Marrow Stromal Cell Culture. Cambridge University Press,

Cambridge UK. 26-42

52. Shi S and Gronthos S. Perivascular niche of postnatal mesenchymal stem cells in

human bone marrow and dental pulp. J Bone and Min Res 2003 18(4): 696-704

53. Delorme B, Ringe J, Gallay N, et al. Specific plasma membrane protein phenotype

of culture-amplified and native human bone marrow mesenchymal stem cells. Blood.

2008; 111: 2631–5.

54. Boiret N, Rapatel C, Boisgard S, et al. CD34_CDw90(Thy-1)_ subset collocated

with mesenchymal progenitors in human normal bone marrow hematon units is

enriched in colony-forming unit megakaryocytes and long-term culture initiating

cells. Exp Hematol. 2003; 31: 1275–83.

55. Deschaseaux F, Charbord P. Human marrow stromal precursors are alpha 1

integrin subunit-positive. J Cell Physiol. 2000; 184: 319–25.

56. Cattoretti G, Schiro R, Orazi A, Soligo D, Colombo M. Bone marrow stroma in

humans: anti nerve growth factor receptor antibodies selectively stain reticular cells

in vivo and in vitro. Blood 1993 81: 1726

57. Crisan M, Yap S, Casteilla L et al. A perivascular origin for mesenchymal stem

cells in multiple human organs. Cell Stem Cell 2008. 3:301–313

92 Chapter 3:

Hyaluronan as a novel marker for rapid selection of connective

tissue progenitors

Tonya Caralla MS1,2, Cynthia Boehm BS1, Vincent Hascall PhD1, George Muschler MD1

1The Department of Biomedical Engineering and the Orthopaedic and Rheumatological

Research Center, The Cleveland Clinic, Cleveland, OH 44195

2 Department of Biomedical Engineering, Case Western Reserve University, Cleveland,

Chapter adapted from paper submitted to the Annals of Biomedical

Engineering

3.1 Abstract

Osteoblastic connective tissue progenitors (CTP-Os) in bone and marrow tissue are an important therapeutic target for bone regeneration. Concentration and selection of CTP-

Os from the heterogenous population of cells in bone marrow remains a challenge due to their low prevalence. This study identifies surface-bound hyaluronan (HA), a component of the in vivo niche for CTP-Os, and evaluates HA as a useful surface marker for positive selection of CTP-Os. Mononuclear cells from bone marrow were labeled and magnetically separated on the basis of hyaluronan binding. HA+++, HA+ and HA- fractions were cultured and assayed for colony formation using a quantitative image-

93 processing system. A mean of 2.7% of cells were retained in the HA+++ fraction and were enriched by 3.4-fold (Range of 95% CI: 2.3-4.8) in CTP-O prevalence when compared to the unselected buffy-coated bone marrow aspirate (BMA). In addition, colonies formed by HA+++ CTPs demonstrated greater proliferation (more cells per colony) and greater alkaline phosphatase expression than CTP colonies derived from unselected BMA. These data demonstrate that one or more subsets of human marrow- derived CTP-Os retain a hyaluronan rich matrix on their surface at the time of harvest.

HA+++ CTPs may offer a useful cell population for regenerative therapies.

3.2 Introduction

Osteoblastic Connective Tissue Progenitors (CTP-Os) provide an attractive cell source for therapeutic applications in regenerative medicine. CTPs are defined as tissue-resident stem or progenitor cells that proliferate to form a colony in vitro and express one or more connective tissue phenotypes. CTP-Os represent the subset of CTPs that demonstrate osteogenic potential.1,2 Bone marrow aspirate is a clinically relevant source of osteogenic cells. However, the prevalence of these cells in bone marrow is low, on average 1 in 20,000 nucleated cells, and this prevalence varies significantly between individuals.3,4 The prevalence of CTP-Os is on the order of 20 fold lower than the prevalence of colony forming hematopoietic progenitors (HPCs) in bone marrow, which have a prevalence of 1 in 1000 nucleated cells.5

Due to the rarity of CTPs in native tissues, many investigators have focused on methods for in vitro expansion of the progeny of CTPs, while retaining osteogenic potential. The

94 terms mesenchymal stem cells or marrow stromal stem cells are used most often to denote these culture expanded populations, which may be derived from the progeny of a pleurality of founding CTPs. Other investigators have focused on methods that enable rapid intraoperative processing of bone marrow to remove red blood cells 6,7, or both red blood cells and non-adherent nucleated cells.8 This paper seeks to further options for intraoperative processing.

CTPs comprise a heterogeneous population of cells in bone marrow that can be found in multiple niches, including the trabecular surface of bone 9, within the highly cellular marrow space, and within a perivascular niche.10-14 Niches for stem cells and progenitors are defined by their anatomical location and also by the local biological environment, which may include the 3D architecture of the niche; cell:cell interactions; paracrine signaling; interaction with local extracellular matrix; oxygen tension; ion gradients; neural inputs; and other metabolic cues.15 These elements of the niche are thought to both define a site that preserves the biological potential of the stem cell or progenitor populations and to regulate number and performance of these cells through influence on the balance of quiescence, proliferation and differentiation.

Hyaluronan (HA) (AKA hyaluronic acid or hyaluronate), is a common but highly regulated component of extracellular matrix. HA is a non-sulfated linear glycosaminoglycan that consists of repeating units of glucuronic acid and N-acetyl- glucosamine. Extracellularly, HA is found as part of the backbone of extracellular matrix structures through proteoglycan binding or attached to the cell surface, forming

95 pericellular coats.16 Chains of HA are synthesized at the plasma membrane on the cytosolic side and extruded through the cell membrane by a hyaluronan synthase (HAS).

Therefore, retention of HA at the cell surface can occur through HAS or through binding to cell HA receptors, notably CD44.

Several researchers have examined histological sections of bone and shown a non- uniform distribution of HA in the bone marrow. In sections of rat tibia, strong HA staining is present in the perivascular area.17 HA staining was also observed surrounding isolated stromal cells in the marrow space.17 Consistent with these results, sections of human bone marrow show staining of HA in stromal extracellular matrix, perivascularly around the marrow sinusoids, and on the endosteal surface of bone.18-20

Upon aspiration or mechanical extraction of cells from the bone marrow, the extracted cells can be expected to retain components of their extracellular matrix niche on their surface. We hypothesized that if HA were a component in the extracellular niche of marrow and bone-derived CTP-Os, HA may represent a distinguishing feature that could provide a method for enrichment of CTP-Os from the heterogeneous population of cells in a fresh bone marrow aspirate. To examine this possibility, we magnetically separated cells from fresh bone marrow aspirates on the basis of hyaluronan retention on the cell surface and assayed for enrichment of CTP-Os.

3.3 Materials and Methods:

3.3.1 Cell Sources

96 Human bone marrow was aspirated from the iliac crest of 11 patients in 2 mL aliquots according to an institutional review board-approved protocol. Canine bone marrow was aspirated from the humerus of 5 donors in 2 mL aliquots according to an approved

IACUC protocol. Two sequential density gradient separations (“buffy coats”) were used to deplete red blood cells. Cells were resuspended in buffer composed of phosphate buffered saline with 2% fetal bovine serum and 1 mM ethylenediaminetetraacetic acid

(EDTA) at 200 million cells/mL.

3.3.2 Magnetic Separation

Cells (200x106/mL) were processed using the EasySepTM Magnetic Separation system

(Stem Cell Technologies #18543) on the basis of HA expression using a biotinylated G1- link protein (hyaluronan binding protein, or HABP) (Calbiochem #385911), which binds strongly and specifically to HA. Cells were sequentially labeled with 200 L/mL of

TM EasySep FC blocker, followed by 20 L/mL of biotinylated HABP at 0.5 mg/mL for 1 hr at room temperature (RT). After removing excess HABP by centrifugation, the

EasySepTM anti-biotin tetrameric antibody complex was added at 200 L/mL and incubated for 15 minutes, followed by EasySepTM magnetic nanobeads at 100 L/mL for a 10 minute incubation. After increasing the total volume to 2.5 mL using buffer solution, the cells were placed in the EasySepTM magnet for 5 minutes allowing magnetized cells to be pulled out of suspension and held against the walls of the container. The nonmagnetized population was decanted. Cells retained in the magnet after 3 sequential resuspension and separation steps were labeled as the enriched HA+++ population. Cells that were not retained after each separation were sent back through the magnet. Any cells

97 retained on a subsequent pass were identified as HA+. Cells that were unbound on all passes through the magnet were defined as HA-.

3.3.3 Hematopoietic Progenitor Cells (HPC) Assay:

To assay the partitioning of HPCs, a standard hematopoietic colony forming unit assay was done in methylcellulose. Cells from each magnetically separated fraction (100,000 nucleated cells in 100 L), as well as the unselected buffy-coated marrow aspirate

(BMA), were combined with 2.5 mL of methylcellulose, 250 L of IMDM serum-free medium, and 150 L of a growth factor cocktail containing 50 ng/mL stem cell factor, 50 ng/mL granulocyte macrophage-colony stimulating factor (CSF), 50 ng/mL granulocyte-

CSF, 10 ng/mL interleukin-3, 3 L/mL erythropoietin, 10 ng/L transforming growth factor FL-1, and 5 ng/L thrombopoietin. Each fraction was plated in duplicate (1 mL per

o well) and cultured in humidified 6 well plates cultured at 37 C at 5% CO2. Erythroid

(CFU-E) and myeloid (CFU-M) colonies were hand counted in each well after 14 days of culture. Prevalence, defined as the number of CFU-M or CFU-E per million cells plated, was calculated and standardized to the prevalence of the BMA.

3.3.4 Colony Forming Unit Assay for CTP-Os

Each fraction was counted with a hemacytometer. BMA, HA+++, HA+ and HA- fractions were cultured in osteogenic medium consisting of -MEM with 10% fetal bovine serum,

1 unit/mL penicillin, 0.1 mg/mL streptomycin, 10-8 M dexamethasone, and 50 g/mL ascorbate (added on the day of use). Each fraction was plated at a density of 500,000 cells

2 o per LabTek chamber (4.2 cm ), and cultured at 37 C at 5% CO2 with medium changes on

98 Days 2 and 3. Day 6 cultures were fixed with 1:1 acetone:methanol for 10 minutes and stained for nuclei (DAPI) and osteoblastic activity (alkaline phosphatase, AP).21

3.3.5 Image Analysis

LabTek culture chambers were scanned using a Spot RTSE 9.0 Monochrome-6 12 bit digital camera (Diagnostic Instruments Inc.) mounted on a Leica DMRXA2 motorized microscope controlled by Metamorph (v6.3) imaging software and analyzed as previously published.21 The ColonyzeTM image analysis software was used to quantitatively identify colonies containing eight or more cells in a cluster (Figure 3.1) using a standardized algorithm.22 CTP Prevalence was defined as the number of CTPs per million cells plated. Each individual colony was characterized based on cell count, colony area (mm2), and AP expression (AP area/cell number).

99 Figure 3.1: The ColonyzeTM software system

3.3.6 Flow Cytometry

A portion of each magnetically separated fraction was reserved for flow cytometric analysis. One million cells in 100 L buffer (PBS with 0.3% bovine serum albumin) were stained for 30 minutes at 4o C and subsequently fixed with 1% paraformaldehyde. All fractions were stained with Streptavidin phycoerythrin (PE) (BD Pharmingen #554061) to detect the biotinylated HABP and with Glycophorin A- fluorescein isothiocyanate

(FITC) (DakoCytomation, F0870) for detection of erthyroid cells, and then run on a

Becton Dickinson LSRII flow cytometer and analyzed with FlowJo software. Secondary

100 antibodies (SA-PE) or isotype controls (FITC anti-mouse IgG2b, BD Pharmingen

553395) were used to determine gating for the experimental samples (N=11).

Seven patients were analyzed with 4 color flow cytometry using HABP-PE, GlyA-FITC,

CD45-PECy7, and CD34-APC. All fractions were stained with Streptavidin PE (BD

Pharmingen #554061) to detect the biotinylated HABP, Glycophorin A-FITC

(DakoCytomation, F0870) for detection of erthyroid cells, CD45-PECy7 (BD

Pharmingen 557748) for detection of committed hematopoietic cells, and CD34-APC

(BD Pharmingen 555824) for detection of hematopoietic stem and progenitor cells.

Samples were run on a BD LSRII and data analysis was performed with FlowJo software.

Fluorescence-minus-one controls using secondary antibodies (SA-PE) or isotype controls

(FITC anti-mouseIgG2b; BD Pharmingen 553395, PECy7 anti-mouse IgG1K; BD

Pharmingen 557646, and APC anti-mouse IgG1K; BD Pharmingen 554681) were utilized to determine gating for the experimental samples.

Gate application on glycophorin A-negative cells eliminated the numerous glycophorin

A-positive red blood cells from analysis. The isotype control for Gylcophorin A was used at 1% positivity (99% negativity) to establish gates. Glycophorin A-negative non- erythroid cells were subsequently analyzed for HA expression to determine the purity and recovery of the magnetic separation. For the 7 patients on which 4 color flow cytometry was performed, CD45 and CD34 expression were analyzed using both HA positive and negative gated cell populations (CD45 and CD34 gates set at 1% negativity using isotype controls).

101

3.3.7 Statistical Analysis

These standardized values were log base 2 transformed to obtain a Gaussian distribution of the response. Means and 95% confidence intervals were calculated. Back transformation provided the geometric mean and the 95% confidence intervals for the geometric mean, providing the relative magnitude of change in the colony level metrics from the control condition. For analysis of cell counts and flow cytometry data, an

ANOVA was used with post-hoc Tukeys test using multiple comparisons (p<0.05).

3.4 Results

3.4.1 Cell and CTP Yield

After magnetic separation, cell counts on each fraction (HA+++, HA+, and HA-) were done.

The mean yield of total nucleated cells after magnetic separation was 79.1 ± 20.7%, indicating that 20.9% of the cells in the starting population were lost during the serial processing steps involved in labeling, washing, and serial resuspension. Of the nucleated cells recovered after separation, 2.7 ± 1.2% of the cells were retained in the HA+++ population, 11.6 ± 5.0% were found in the HA+ population, and 85.7 ± 5.1% were HA-.

The range of CTP enrichment observed after magnetic separation is illustrated in Figure

3.2. The total number of CTPs was calculated by multiplying the prevalence of CTPs in

102 the initial BMA sample by the total number of nucleated cells in the starting sample. The total number of CTPs in the HA+++, HA+, and HA- populations was calculated using the same method. Figure 3.3 illustrates the partitioning of nucleated cells and CTPs in each fraction after magnetic separation. The HA+++ fraction only contained 2.0% of the nucleated cells, but 8.3% of all CTPs. The HA- fraction contained the majority of nucleated cells (66.7%) and 27.4% of the CTPs.

Figure 3.2: CTPs are enriched in the HA+++ fraction. Each column contains representative ColonyzeTM images for 4 patients after magnetic separation. CTP

103 prevalence of each fraction and fold change in prevalence of each fraction over BMA are listed. Images: Top row: unselected buffy coated marrow aspirate (BMA), middle row: HA+++ fraction, bottom row: HA- fraction.

The loss of colony forming CTPs during processing was greater than the loss of cells in general. These calculations suggest that 59.5% of the CTPs in the initial sample were lost during the multiple processing steps (e.g. via adherence to surfaces and/or greater buoyancy following centrifugation), in contrast to a 20.9% loss of nucleated cells.

Figure 3.3: Cell and CTP partitioning after magnetic separation. Bars are standard error. N=11. Cell counts: HA- is significantly different than HA+++ (p<0.0001) and HA+ (p<0.0001). CTP counts: HA- is significantly different than HA+ (p=0.0215).

3.4.2 Purity and Recovery

Flow cytometric analysis was used to assess the purity of the HA+++ fraction and the recovery of HA-positive cells after magnetic separation. Erythroid cells were labeled with a glycophorin-A antibody to provide a method for eliminating erythroid cells from

104 analysis. The non-erythroid nucleated cell fraction was selected using a Glycophorin-A negative gate, and the HA-positive cells were subsequently quantified. Figure 3.4 illustrates the results from flow cytometry.

Figure 3.4: Flow cytometric analysis of each fraction after magnetic separation. Non- erythroid cells are gated on the basis of glycophorin-A negativity, then analyzed for HA positivity to calculate purity and recovery. Negative controls (blue) are overlayed with an HA+++ sample to illustrate gating techniques. N=11.

Magnetic separation was effective in isolating cells that retained or expressed HA on their surface. The mean purity of the HA+++ fraction was 54.3 ± 18.6% (n =11). A purity of 100% would indicate that all the cells isolated in the HA+++ fraction are positive for

HA.

105 Magnetic separation was effective in recovering cells with HA on their surface. The mean recovery was 78.2 ± 12.5% in the HA+++ fraction alone. The sum of both HA- positive fractions (both HA+++ and HA+) contained 92.7 ± 8.3% of all cells presenting HA on their surface. A recovery of 100% would indicate that there are no HA-positive cells in fractions other than the HA-positive fraction(s). Therefore, less than 8% of all cells found to stain positive for HA on their surface were recovered in the HA- fraction.

3.4.3 Four color flow cytometric analysis

Seven patients were stained with HABP-PE, GlyA-FITC, CD45-PECy7, and CD34-APC.

To eliminate the numerous glycophorin A-positive red blood cells from analysis, a gate on glycophorin A- cells was applied. These non-erythroid cells were evaluated for HA- positive and HA-negative cells. Both HA-positive and HA-negative cells were further analyzed for CD45 and CD34 expression to evaluate the types of cells that may co- separate with CTP-Os. This was repeated for each cell fraction obtained by magnetic separation (HA+, HA+++, HA- fractions) as well as the unselected BMA. Table 3.1 shows the mean and standard deviation for each of these groups.

Percent of cells that Percent of non- Percent of HA+ non- Percent of HA+ non- Percent of HA- non- Percent of HA- non- are non-erythroid erythroid cells that erythroid cells that erythroid cells that erythroid cells that erythroid cells that Sample (glycophorin A-) are HA+ are CD45+ are CD34+ are CD45+ are CD34+ % of total sample % of GlyA- cells % of GlyA- HA+ cells % of GlyA- HA+ cells % of GlyA- HA- cells % of GlyA- HA- cells Buffy 53.4% ± 24.2% 10.0% ± 8.4% 68.9% ± 20.0% 1.2% ± 2.0% 43.2% ± 22.8% 0.8% ± 1.5% HA+ 58.9% ± 30.8% 10.9% ± 5.6% 57.5% ± 35.3% 0.6% ± 1.0% 39.1% ± 24.6% 1.1% ± 1.8% HA+++ 95.4% ± 2.9% 50.5% ± 19.9% 30.0% ± 18.4% 0.3% ± 0.5% 36.5% ± 17.9% 0.2% ± 0.3% HA- 52.1% ± 20.4% 3.7% ± 4.0% 68.8% ± 19.7% 0.2% ± 0.3% 28.0% ± 16.4% 0.1% ± 0.1%

Table 3.1: Analysis of 4 color flow cytometry on each cell fraction. The mean values and standard deviation is shown for the 7 patients.

For the 4 color flow cytometric analysis, significant differences were found in the percent of non-erythroid (glyA-) cells partitioning to each fraction. The HA+++ fraction contained significantly more non-erythroid cells than the BMA (p=0.0058), HA+ (p=0.0193), and

106 HA- (p=0.0043) fractions. Glycophorin A-positive RBCs partitioned primarily to the

HA- and the HA+ fractions. As expected, the non-erythroid HA+++ cell fraction contained more HA-positive cells than the BMA, HA+, and HA- non erythroid fractions (p<0.0001).

CD34 is a protein expressed on hematopoietic stem and progenitor cells. The detection of

CD34+ cells was low in all fractions (<1.5%). Of the non-erythroid HA-positive cells, no difference was found in the expression of CD34+ cells between any of the fractions.

CD45+ cells are typically mature hematopoietic cells. Of the non-erythroid HA-positive cells, the HA+++ fraction contained significantly less CD45+ cells than the BMA

(p=0.0004), HA+ (p=0.0101), and HA- (p=0.0004) fractions. Of the non-erythroid HA- negative cells, no difference was found in the expression of CD34+ cells or CD45+ cells between any of the fractions.

3.4.4 Selection of Colony Forming HPCs

Hematopoietic progenitor cells were negatively selected (depleted) in cells presenting HA on their surface. The relative prevalence of HPCs in the HA+++ fraction were 0.03 for

CFU-M (a 33 fold reduction from the BMA) and 0.11 for CFU-E (a 9 fold reduction from the BMA). (Figure 3.5). Both myeloid and erythroid progenitors partition to the HA- fractions, which show similar prevalence to the unselected marrow.

107

Figure 3.5: Prevalence of myeloid and erythroid progenitors after magnetic separation. The HA+++ fraction is significantly depleted in both myeloid and erythroid progenitors.

108 3.4.5 Selection of Colony Forming CTPs

As illustrated in Figure 3.6, the HA+++ fraction was significantly enriched in colony forming CTPs, with an average of 3.4-fold enrichment over the BMA, and a 10.0-fold enrichment over the HA- fraction. The HA+ and HA- fractions were significantly depleted in progenitors (0.50 and 0.35, respectively) compared to the BMA.

Figure 3.6: Quantification of CTP prevalence demonstrates that the HA+++ fraction is significantly enriched in CTPs when compared to the unselected bone marrow aspirate control (BMA). HA+ and HA- fractions are significantly depleted in progenitors.

3.4.6 Biological Performance of Colonies Derived form CTPs Isolated in the HA+++,

HA+, and HA-

3.4.6.1 Proliferation:

The progeny of CTPs in the HA+++ population demonstrated a greater proliferation rate than CTPs in unprocessed marrow or in the HA- fraction (figure 3.7). The colonies derived from CTPs in the HA+++ population contained a mean number of cells per colony

109 that was 1.8-fold greater than the colonies derived from unprocessed BMA. The mean cells per colony in the HA+ population was not different than unselected marrow, while the HA- fraction had significantly fewer cells per colony than the BMA (0.71).

3.4.6.2 Alkaline Phosphatase Activity:

Colonies derived from CTPs in the HA+++ fraction showed 2.7-fold greater AP expression (area per cell) than colonies derived from unprocessed BMA, suggesting that the progeny of HA+++ CTPs can quickly more exhibit a CTP-O phenotype than colonies derived from HA- CTPs (figure 3.7).

110

Figure 3.7: CTP-derived colonies in the HA+++ fraction have significantly more cells per colony, indicative of increased proliferation, when compared to the unselected BMA. Colonies formed from HA- CTPs have significantly less cells per colony. Colonies derived from HA+++ CTPs have significantly increased AP expression over the BMA, indicative of osteoblastic differentiation.

111 3.4.6.3 Retained versus Newly Synthesized Hyaluronan:

An assumption in this work is that freshly isolated CTPs will retain a pericellular coat of hyaluronan from their in vivo niche. However, it is also possible that the cell-associated

HA could result from new synthesis of HA. Due to the fact that the HAS enzyme is inactivated at 4oC, it is possible to prevent new HA synthesis by processing cells at 4oC.

Therefore, two canine aspirates were processed in parallel, one at room temperature and one at 4oC, following the normal protocol for magnetic labeling and HA-based selection

(N=5). The percent of cells partitioning to the HA+++ fraction at room temperature (RT) was not different than with processing on ice (1.13 ± 0.33% vs. 1.72 ± 1.24). CTP prevalence in the HA+++ fraction was not significantly different between RT and ice samples. At RT the HA+++ fraction was enriched in CTPs an average of 1.29-fold over the BMA; and when processed on ice, the enrichment in the HA+++ fraction was 1.37-fold.

Using both RT and ice processing, the HA+ and HA- fractions were both significantly decreased in CTPs when compared to the BMA. This suggests that HA is retained on the cell and not synthesized after aspiration.

Cell and CTP partitioning using canine marrow was similar to human marrow. The mean number of cells partitioning to the HA+++ fraction using canine bone marrow was analogous to that obtained using human bone marrow (canine: 1.13 ± 0.33% vs. human

2.7 ± 1.2%). CTP prevalence also followed the same fractionation after magnetic separation using canine marrow: an increase in CTPs in the HA+++ fraction and decrease in CTPs in the HA+ and HA- fractions.

112 3.5 Discussion

These data demonstrate that at least one subset of the heterogeneous population of marrow derived CTPs is enriched by selecting for cells that present HA on their surface.

In comparison to the unselected marrow, these cells are significantly more proliferative and more readily differentiate to an osteoblastic phenotype in vitro. In contrast, both myeloid and erythroid hematopoietic progenitor cells do not segregate with CTPs and remain in the HA- fraction.

The enrichment of osteogenic CTPs based on HA, an extracellular matrix molecule, suggests that HA represents one component of the niche surrounding at least one osteogenic progenitor cell population in human bone marrow. These data demonstrate that CTPs retain this coat of HA rather than synthesize it after aspiration. There was no difference in mean CTP enrichment after magnetic separation between samples processed on ice (during which the HAS enzymes are inactivated) and at room temperature.

The anatomic origin of the HA+++ CTPs isolated in this study is not known, however, histological analyses show that HA is observed in the ECM around two subsets of cells in human marrow. Published histology shows sporadic HA staining of stromal cells in the marrow space and strong staining of HA in the perivascular area.17-20 Evidence supports a perivascular niche for at least one subset of osteogenic cells in marrow.12,13 Stro-1 has been used to isolate colony forming cells, and CD146 is a pericyte marker that is also expressed on culture expanded MSCs.11 Stro-1+/CD146+ cells are found lining blood vessels in human bone marrow and dental pulp.10 Crisan et al. did fluorescence activated

113 cell sorting on cells from multiple tissues, including bone marrow. CD56- CD45-

CD146+ CD34- cells differentiated along the chondrogenic, adipogenic, and osteogenic lineages in vitro, and formed bony nodules in vivo when implanted into immunocompromised mice.14

While these data suggest that at least one subset of CTPs resides in a niche that is characterized by HA, it also suggests that this is not true for all marrow-derived CTPs.

Colony forming CTPs are found in both the HA+++ and the HA- population. While the more rapidly proliferative and more osteogenic HA+++ fraction is of greatest interest in the domain of fracture repair, remodeling and bone regeneration, many colony forming

CTPs are negative for HA. Among the CTPs isolated in the processed fractions, the total number of HA- CTPs was 3.3 fold greater than HA+++ CTPs. Characterization of the differences between HA+++ and HA- CTP populations will require further refinement of methods for separation.

The magnetic cell separation methods used in these experiments can be readily adapted for intraoperative enrichment of osteogenic CTPs from autogenous bone marrow prior to transplantation. Using current methods, the total time for labeling (1.5 hrs), and separation (15 min) pushes the limit of clinical practicality. However, current methods can be simplified to reduce labeling time, and readily scaled for processing up to 2x109 cells.

114 The methods of magnetic separation used in this study are as yet imperfect. The analysis of CTP yield in comparison to the starting sample suggests that 59% of all CTPs in the starting sample were lost during processing. The true ratio of HA+++ to HA- CTPs cannot be determined until the nature and fate of CTPs that were lost during processing has been determined. The most likely cause of CTP loss during processing is adherence to surfaces during processing. CTPs are characterized by a tendency to attach readily to tissue culture plastic and other surfaces. The extended number of steps and incubation periods associated with current methods provide abundant opportunity for cell loss due to attachment. A loss of viability is also a possible cause, but is not suspected. It is also possible that the labeling procedures and processing resulted in a change in cell state that diminished adherence or proliferative capacity, reducing colony forming efficiency.

However, the robust colony formation of the adherent CTPs that complete processing does not suggest a change in viability or performance related to processing.

The observations in this study give rise to several important questions with clinical relevance. The fact that HA+++ CTPs demonstrate significantly more cells per colony

(indicative of increased proliferation), and also increased alkaline phosphatase compared to the BMA, suggest that this subset of CTPs may be preferred for tissue engineering applications.

These data also give rise to biologically relevant questions regarding the epidemiology and kinetics of CTPs in bone marrow. What is the anatomic location of the HA+++ CTPs as well as the more numerous HA- CTP population, and are these subsets of progenitor

115 cells related? How do these populations differ in function and biological potential? What is the role of HA in the niche around the HA+++ CTP, and what other components are present?

3.6 Acknowledgments:

The authors would like to acknowledge Richard Rozic (imaging), Jason Bryan

(ColonyzeTM software development), and Sean Fleury (image review). Funding was provided by: Ohio Department of Development, NIH R01 AR42998 “Optimizing Bone

Marrow as a Bone Graft”.

3.7 References:

1. Muschler G, Midura R. 2002. Connective tissue progenitors: practical concepts for clinical applications. Clin Orthop 395: 66-80

2. Muschler G, Nakamoto C, Griffith L. 2004 Engineering principles of clinical cell- based tissue engineering. J Bone Joint Surg 86A(7): 1541-1558

3. Muschler G, Boehm C, Easley K. 1997 Aspiration to obtain osteoblast progenitor cells from human bone marrow: the influence of aspiration volume. J Bone Joint Surg Am 79:

1699-1709

4. Majors A, Boehm C, Nitto H, et al. 1997 Characterization of human bone marrow stromal cells with respect to osteoblastic differentiation. J Orthop Res 15: 546-557

5. Verfaillie C, Blakolmer K, McGlave P. 1990 Purified primitive human hematopoietic progenitor cells with long term in vitro repopulating capacity adhere selectively to irradiated bone marrow stroma. J Exp Med 172: 509-520

116 6. Connolly J. 1995 Injectable bone marrow preparations to stimulate osteogenic repair.

Clin Orthop 313: 8-18

7. Hernigou P, Poignard A, Manicom O, Mathieu G, Rouard H. 2005 The use of percutaneous autologous bone marrow transplantation in nonunion and avascular necrosis of bone. J Bone Joint Surg Br 87(7): 896-902

8. Muschler G, Matsukura Y, Nitto H, Boehm C, Valdevit A, Kambic H, Davros W,

Easley K, Powell K. 2005 Selective retention of bone marrow-derived cells to enhance spinal fusion. Clin Orthop 432: 242-51

9. Patt HM, Maloney MA. 1972 Relationship of bone marrow cellularity and proliferative activity: a local regulatory mechanism. Cell Tissue Kinet 5(4): 303-309

10. Shi S, Gronthos S. 2003 Perivascular niche of postnatal mesenchymal stem cells in human bone marrow and dental pulp. J Bone Miner Res 18: 696-704

11. Sorrentino A, Ferracin M, Valtieri M et al. 2008 Isolation and characterization of

CD146+ multipotent mesenchymal stromal cells. Exp Hematol 36: 1035-1046

12. Kolf C, Cho E, Tuan R. 2007 Biology of adult mesenchymal stem cells: regulation of niche, self-renewal and differentiation. Arthritis Res and Therapy 9: 204-214

13. da Silva Meirelles L, Chagastelles P, Nardi N. 2003 Mesenchymal stem cells reside in virtually all post-natal organs and tissues. J Cell Sci 119: 2204-2213

14. Crisan M, Yap S, Casteilla L et al. 2008 A perivascular origin for mesenchymal stem cells in multiple human organs. Cell Stem Cell 3: 301-313

15. Scadden D. 2006 The stem cell niche as an entity of action. Nature 441: 1075-1079

16. Wang A, Hascall V. 2004 Hyaluronan structures synthesized by rat mesangial cells in response to hyperglycemia induce monocyte binding. J Biol Chem 279: 10279-10285

117 17. Midura R, Su S, Morcuende J et al. 2003 Parathyroid hormone rapidly stimulates hyaluronan synthesis by periosteal osteoblasts in the tibial diaphysis of the growing rat. J

Biol Chem 278: 51462-51468

18. Avidgor A, Goichberg P, et al. 2004 CD44 and hyaluronic acid cooperate with SDF-1 in the trafficking of human stem/progenitor cells to bone marrow. Blood 103(8): 2981-

2989

19. Sundstrom G, Lofvenberg E, Hassan I et al. 2002 Localisation and distribution of hyaluronan in normal bone marrow matrix: a novel method to evaluate impending fibrosis? Eur J Haematol 68: 194-202

20. Sundstrom G, Dahl I, Hultdin M et al. 2005 Bone marrow hyaluronan distribution in patients with acute myeloid leukemia Medical Oncology 22: 71-78

21. Villarruel S, Boehm C, Pennington M et al 2008 The effect of oxygen tension on the in vitro assay of human osteoblastic connective tissue progenitor cells J Orthop Res

26(10): 1390-1397

22. Powell K, Nakamoto C, Villarruel S et al. 2007 Quantitative image analysis of connective tissue progenitors. Anal Quant Cytol Histol 29(2): 112-121

118 Chapter 4:

A Model of Retention of Magnetized Cells in the EasySep Magnetic

Separation System

Abstract:

Connective tissue progenitors (CTPs) are a valuable cell source for regenerative therapies.

These cells are present in bone marrow aspirates, however, CTPs among nucleated cells in marrow aspirates have a low prevalence. This rarity highlights the need for cell separation technologies to select these cells from the heterogeneous mix of cell types present in a marrow aspirate. Using magnetic separation, new markers can be evaluated for specificity to CTPs to enrich this rare population of cells and eliminate the majority of other competing, non-osteogenic cell types.

Bone marrow aspirates are labeled with an antibody and magnetic bead and place into the magnet, which retains positively labeled cells while the non-magnetically labeled cells are removed. In order to ensure the magnetically tagged cells are being retained in the magnet, a mathematical model was created to determine the velocity experienced by the cells and the retention time necessary to ensure recovery of these tagged cells.

119 4.1 Introduction

Cell separation methods provide powerful tools for discriminating between diverse cell types and investigating cells for various surface markers. Within a mixture of cells, a specific cell type can be identified using previously established markers; furthermore, these specific cells can be isolated to produce a purified population. Alternatively, a single known cell type can be screened for a new potential marker. Examples of both types of separation strategies utilize bone marrow as its cell source.

Human bone marrow is added to bone graft implants to regenerate bone damaged by tumor, infection, trauma, or severe fractures. The osteogenic cells resident in bone marrow, termed connective tissue progenitors, or CTPs, are the desired cell type for implantation into these grafts. However, CTPs among nucleated cells in marrow aspirates are very rare, on average 1 in 20,000 cells1,2, and this rarity highlights the need for separation technologies to select these cells from the heterogeneous mix of cell types present in a marrow aspirate. Using magnetic separation, new markers can be evaluated for specificity to CTPs to enrich this rare population of cells and eliminate the majority of other competing, non-osteogenic cell types.

Another therapy utilizing aspirated bone marrow is the transplantation of hematopoietic stem cells (HSC) and bone marrow in the treatment of hematologic and lymphoid cancers.

HSCs have a higher baseline prevalence in marrow, around 1 in every 200 cells, and are responsible for the maintenance of all myeloid and lymphoid lineages. It has been established that HSCs express the cell surface marker CD34, and current clinical

120 therapies to treat deficiencies or cancers of immune cells (lymphoma, myeloma, and amyloidosis) use antibodies to CD34 to select HSCs for transplantation using cell separation strategies.3,4

Currently available cell separation technologies include density gradient separation, flow cytometry, and magnetic separation. Density gradient separation uses inherent differences in cell density to separate out nucleated cells from red blood cells in a cell suspension.

However, density gradient separation is limited by inefficiencies in recovery of cells and doesn‟t provide any selection within the mononuclear cell fraction. It is essentially a technology for concentrating mononuclear cells by removing the large red cell fraction and pelleting the remaining cells in a small volume.

Flow cytometry uses differences in cell size and granularity to separate unlabeled cells

(lymphocytes, monocytes, and neutrophils form distinct discrete clusters), and cells can also be tagged with fluorescent antibodies for any cell surface marker and separated based on the intensity of the fluorophore. This is a powerful tool for cell separation and cells can be separated into multiple fractions based on clustering or differences in fluorescence intensity.

Magnetic separation uses the technique of coupling magnetic beads to antibodies for cell surface markers. Cells tagged in this way and placed into an external magnetic field will move toward the magnet wall and be retained by the magnetic field, while unlabeled cells in the mixture will remain in suspension. This “negative” unlabeled fraction can be

121 removed and a purified population remains. Magnetic separation addresses some of the limitations of flow cytometry, namely, higher throughput and the elimination of the high shear forces present in the cytometer. Commercially available magnetic separation systems can manage 100 million to 1 billion cells in a small permanent magnet.

Disadvantages of magnetic separation include a binary selection, as well as the same animal-derived antibody biocompatibility issues that affect the clinical use of flow cytometry. Both systems can select a wide range of cell types using customizable biotin beads.

Magnetic separation systems offer the advantages of ease-of-use, increased viability of

“fragile” cells, and convenient, inexpensive equipment that can handle separation of

200x106 cells in a single tube in a small permanent magnet. Of the commercially available magnetic separation systems, the EasySep system from Stem Cell Technologies was selected as the magnetic separation system for screening cell surface markers on cells from a fresh human bone marrow aspirate. This system offers all of the advantages listed above, and the additional advantage of a column-free separation.

In order to ensure the magnetically tagged cells are being retained in the EasySepTM magnet, a mathematical model was created to determine the velocity experienced by the cells and the retention time necessary to ensure recovery of these tagged cells.

4.2 Materials and Methods:

4.2.1 Magnetic Labeling in the EasySepTM System

122 The magnetic beads used to label cells in the EasySepTM system are dextran-coated magnetite. Magnetite (Fe3O4, or FeO*Fe2O3) is a naturally occurring mineral that is ferrimagnetic. However, it can be processed into small particles that behave paramagnetically but retain their large susceptibility, and thus the beads used in the

EasySep system are defined as superparamagnetic particles. A diagram of a cell tagged with these magnetic particles is presented in figure 4.1. In this example, hyaluronan, an extracellular matrix component, is used as the selection marker. Fresh human bone marrow aspirates were stained sequentially with a biotinylated hyaluronic acid binding protein, followed by the EasySep anti-biotin tetrameric antibody complex. Finally, the magnetic beads are added and the cells are fully labeled and ready for placement in the magnet.

Figure 4.1: Example of the staining sequence with EasySepTM reagents using hyaluronan as a marker for positive selection.

4.2.2 Geometry of the EasySep Magnet

123 The EasySep magnet, pictured in figure 4.2a, is a small permanent magnet that accommodates a 5mL polystyrene falcon tube (diameter = 11mm) containing the magnetically labeled cell suspension. Using iron filings to uncover the magnetic field lines yielded the images in figure 4.2b-c. Magnetic field lines provide a visual representation of the magnetic field where the field line density is proportional to the magnitude of the magnetic field.

As seen in figure 4.2b-c, the EasySep magnet utilizes a quadrupole geometry, indicating that there are 4 permanent magnets oriented at 0o, 90o, 180o, and 270o. A quadrupole magnet offers compact design while generating high magnetic fields and high gradients.

An ideal quadrupole field produces only a centrifugal force field, as B is dependent on radial position r only.7

A B C

Figure 4.2: A) Top view of the EasySepTM magnet. B-C) Iron filings laid over the top of the EasySepTM magnet illustrate the magnetic field lines between the poles of the magnets. B. Four poles can be seen at 0o, 90o, 180o, and 270o positions. C. A closer view of the field lines between 2 poles.

4.2.3 Magnetic Flux Density of the EasySep Magnet

The magnetic flux density B in the bore of the EasySep magnet was obtained experimentally using a Sypris 5100 Series Hall Effect Gaussmeter/Teslameter. The principle of operation of this equipment is based on the Hall Effect, which states that when current flows through a conductor in a magnetic field a charge separation builds on

124 the conductor, producing a measurable voltage. The Gaussmeter measures the voltage produced by this charge separation, which is proportional to the magnetic flux density B:

B I V   d H ne where VH is the Hall Effect voltage produced, I is the current across the plate length, B is the magnetic flux density, d is the depth of the plate, e is the electron charge, and n is the charge carrier density of the carrier electrons. The EasySep magnet displayed a linear relationship between the measured magnetic flux density B and radial position r. The graph is illustrated in figure 4.3. Both sets of data are displayed, as well as the trendline with R2 value. The R2 value is 0.9999, showing strong data correlation.

EasySep Magnetic Field Measurement (n=2)

0.8

0.7 y = 148.12x 2 0.6 R = 0.9999

0.5

0.4

0.3

0.2 MagneticFluxDensity (T)B 0.1

0.0 0 0.0005 0.001 0.0015 0.002 0.0025 0.003 0.0035 0.004 0.0045 0.005 Radial Distance r from Center of Magnet (m)

Figure 4.3: Measurement of the magnetic flux density B as a function of radial position r.

4.2.4 Testing with standardized Micromod beads

125 Micromer-M beads (08-00-603) from Micromod were obtained to test the model parameters. These standardized beads are made from magnetite in polymer matrix and are fairly monodisperse with a diameter of 6.48 ± 0.62 m as measured using a Coulter counter. These beads are ideal for testing the mathematical model because their size approximates that of bone marrow mononuclear cells, they contain the same magnetic carrier as the EasySepTM beads, and since they are monodisperse and contain the same amount of magnetite per bead, the results are predictable and reproducible. In addition, the magnetophoretic mobility of these beads has been measured and the mean magnetophoretic mobility m is 7.44x10-13 m3/(T*A*s), which is comparable to the mean magnetophoretic mobility of the labeled cells.

One milliliter of well-mixed beads in aqueous buffer in a 5mL polystyrene falcon tube were placed in the EasySepTM magnet, and the beads were left for varying amounts of time in the magnet (time intervals tested: 5, 10, 15, 25, 35, 45, and 60 seconds). When the residence time in the magnet ended, the “negative” solution was pipetted out, and counts were performed on the retained (positive) and non-retained (negative) fractions using an automated Vi-Cell counter. The percentage of retained beads was calculated and compared to the percentage of retained beads predicted by the mathematical model.

4.2.5 Generation of HA+++ cells for CTV analysis

Cells (200x106/mL) were processed using the EasySepTM Magnetic Separation system

(Stem Cell Technologies #18543) on the basis of HA expression using a biotinylated G1- link protein (hyaluronan binding protein, or HABP) (Calbiochem #385911), which binds

126 strongly and specifically to HA. Cells were sequentially labeled with 200 L/mL of

TM EasySep FC blocker, followed by 20 L/mL of biotinylated HABP at 0.5 mg/mL for 1 hr at room temperature (RT). After removing excess HABP by centrifugation, the

4.2.6 Cell Tracking Velocimetry

Cell tracking velocimetry (CTV) generates a magnetic field that contains a defined region with constant properties. The magnet used in the CTV apparatus contains a region of constant energy density gradient SM where the imaging and analysis occurs. The protocol and experimental setup of CTV have been described in detail elsewhere.10-12

Briefly, the HA+++ cell suspension is injected into a channel surrounded by a dipole magnet. The magnet is oriented perpendicular to gravity forces. After injection, there is a

127 time lapse before measurement in the field to negate the effects of convection generated by the injection. Magnetically tagged cells will migrate across the channel toward the magnet, while untagged cells will not move, or will sediment in a direction perpendicular to the magnetic force. A camera captures the magnetically induced motion of the cells during a user-defined time frame. A series of photos records the spatial position of the particle over time, and a 2D tracking module determines the location of each cell and its change in position.12 The output of the system is the velocity of the HA+++ cells and their magnetophoretic mobility.

4.3 Theory and Calculations:

4.3.1 Derivation of model

Starting with a force balance on the labeled cell in the magnetic field, the cell experiences a magnetic force FM opposed by a drag force FD in the x-direction, as well as a gravity force FG opposed by buoyancy FB and drag forces FD in the y-direction (figure 4.4).

FB FD

FM FD

F G

Figure 4.4: Force Diagram on the magnetized cell where FM= Magnetic Force, FD= Drag Force, FB=Buoyant Force and FG=Gravity Force

128 From Newton‟s Second Law, that the sum of all forces in a given direction is equal to the mass of the magnetized cell:label ensemble (referred to afterward as the magnetized cell) multiplied by the acceleration.

 F  ma

For small particles in liquid viscous media, the inertial term can be neglected so that ma =

0. Because the magnetized cell reaches equilibrium in the solution essentially instantaneously, it travels at a velocity relative to the ratio of the magnetic force to the drag force.8 For justification of this assumption see Appendix A.

In this problem, magnetic forces and drag forces operate in the x-direction, while sedimentation operates in the y-direction. If sedimentation << magnetic forces (Appendix

B) then sedimentation can be neglected and

 F  FM  FD or FM  FD

The magnetic force FM is defined as

FM  MVB where M is the volume magnetization, V is the total volume of magnetic carrier, andB is the gradient of the magnetic field intensity. This equation is adapted from the

M equation relating susceptibility to magnetization (   ) and magnetic force H

F  VHB .7

In cylindrical coordinates, the gradient of B is defined as:

129 dB 1 dB dB B    dr r d dz

dB In the axial direction (z) the magnetic field is constant so that  0 . Assuming an ideal dz

dB quadrupole magnet, the change in is negligible so that  0 . 5,7 The relationship d between the radial distance r and the magnetic field intensity B is linear in the quadrupole magnetic field (figure 4.5) so that:

dB B   148.12 dr where 148.12 (T/m) is the slope of the line B=148.12r.

The drag force FD is defined by the Stokes formula for viscous media with creeping flow, or a Reynolds number NRe <0.1.

FD  6Rv where  is the viscosity of the buffer, R is the magnetized cell radius, and v is the velocity.

Note that R is strictly defined as the radius of the magnetized cell (the cell plus the magnetic bead label), but given that the diameter of an average cell (9.7x10-6 m) is an order of magnitude larger than that of the bead (1.6x10-7m), it is assumed that use of the cell radius alone is sufficient. In addition, note that a variety of cell types are present in bone marrow and therefore a range of cell diameters are observed between 5-17m with a mean of 9.7m. A representative average cell radius of 4.85m is used in these

130 calculations. See figure 4.5 for the distribution of cell size in a representative bone marrow aspirate, as measured by an automated Vi-Cell counter.

Figure 4.5: Histogram of bone marrow aspirate obtained by Vi-Cell counter analysis. A representative sample of the wide range of cell diameters present in a fresh bone marrow aspirate. The average cell diameter of 9.7m was used in this model.

The Reynolds number is given as NRe:

Dv N  Re  where D is the diameter of tube, v is the velocity,  is the density of the magnetized cell, and  is the viscosity. Here the velocity v≈0 for a stationary fluid, so that NRe ≈0. This satisfies the requirement for NRe <0.1 for use of the Stokes equation.

Setting FM=FD, this equation describes the motion of the cells in the magnetic field.

MVB  6Rv

The unknown velocity v is the desired variable. However, the total volume of magnetic carrier V is unknown as well as the number of beads per labeled cell.

131 4.3.2 Relating model equation to magnetophoretic mobility

Using the cell tracking velocimetry (CTV) system, the magnetophoretic mobility m of the stained cells can be measured. An alternate solution to this problem relates the unknown

EasySepTM system variables to the well-characterized CTV system.

Variables that remain constant between the EasySep and CTV systems include:

o Radius of labeled cell R

o Volume of magnetic carrier V

o Viscosity of aqueous buffer 

o Constant term 6

Variables that change between the EasySep and CTV system are the characteristics of the magnet and the unknown velocities:

o Magnetization M because magnetization is a function of B

o Gradient of the magnetic field intensity B

o Velocity v

Setting the constant terms aside,

6R MB  V v and defining the subscript 1 for the CTV system and 2 for the EasySepTM system:

M B M B 1 1  2 2 v1 v2

Solving for v2 (the velocity of the cells in the EasySep system):

M 2B2v1 v2  M1B1

132 The CTV system reports the magnetophoretic mobility as well as the velocity of magnetized cells. The magnetophoretic mobility m is defined as:

v m  S M

B 2 with S M  2o where SM is the magnetophoretic driving force, and o is the magnetic permeability of free space.

Plugging in m1 and SM1 for CTV,

M 2B2 v2  m1S M1 M1B1

Since the CTV system operates in the saturated area of the curve, M1=Msat of magnetite.

However, M2=M(B) for magnetite in the range experienced in the magnetic field in the

EasySep system.

Velocity v is the change in distance per unit time, so:

dr v  dt where r is position or distance. Plugging in for M1SAT and v2:

dr B2 M 2 (B) v2   m1SM1 dt B1 M1SAT

Table 4.1 provides the parameter values used in this equation.

133 Table 1: Calculation of Velocity of Magnetized Cells

dr B2 M 2 (Bo ) v2   m1SM1 dt B1 M1SAT Symbol Value Units Description Source 3 m1 1.04E-12 m /(T*A*s) CTV magnetophoretic mobility Measured in CTV experiment 2 SM1 1.40E+08 T*A/m Magnetophoretic driving force Constant in CTV system B2 148.12 T/m EasySep gradient of B Experiment - slope of trendline (Figure 5) B1 140.3 T/m CTV gradient of B Constant in CTV system M1SAT 342435 A/m Magnetization (saturated) in CTV Calculated from curve with B=140.3 T/m (Figure 8)

M2(B) equation A/m EasySep magnetization Equation from curve (Figure 8) B(r ) 148.12*r T EasySep field intensity Experiment - trendline (Figure 5) v2 variable of interest m/s Velocity in EasySep system Equation

Table 4.1: Parameters used for the calculation of the velocity of the magnetized cells

The equation relating magnetization M to flux density B for superparamagnetic magnetite particles is given by:9

9.152106 * B M  1 27.304* B  0.92289* B 2

In a quadrupole magnet, flux density B is linearly dependent on position r. B was measured with a Hall Effect Gaussmeter, and the trendline was calculated and averaged to give the relationship (see Figure 3):

B  148.12r

Substituting in for B, magnetization M for the EasySep system becomes:

1.361109 *r M  1 4060*r 137.2*r 2

Using the equation for v2 and plugging in for M2:

9 dr B2 1 1.36110 *r v2   m1SM1 2 dt B1 M1SAT 1 4060*r 137.2*r

Now there is essentially some constant number that is composed of

B2 1 m1SM1 B1 M1SAT

And an equation dependent on radial position:

134 1.361109 *r

1 4060*r 137.2*r 2

In order to determine the dependence of position r with respect to time t, we separate variables and integrate.

1 4060r 137.2r 2 m S B dr  1 M1 2 dt  9  1.36110 r M1SAT B1

The left side can be simplified to:

7.3481010 m S B  (  2.983106 1.01107 r)dr =  1 M1 2 dt r M1SAT B1

Upon Integrating:

r 2 2 m S B t (7.348 1010 *ln(r)  2.983106 * r  5.05108 *r )  1 M1 2 t M B 0 r1 1SAT 1

Evaluating the limits:

10 r2 6 8 2 2 m1SM1 B2 (7.348 10 *ln( )  2.98310 *(r2  r1 )  5.0510 *(r2  r1 )  t r1 M1SAT B1

This is the final equation relating the radial position r to the time required for the particle to reach the wall of the tube and be retained.

4.4 Results:

4.4.1 Standardized Micromod Beads

Micromer-M beads (#08-00-603) from Micromod were obtained to test the model parameters. Using the measured magnetophoretic mobility (7.44x10-13 m3/(T*A*s)) and the EasySep and CTV parameters, the required retention time for each radial position r was calculated using the model equation:

135 10 r2 6 8 2 2 m1SM1 B2 (7.348 10 *ln( )  2.98310 *(r2  r1 )  5.0510 *(r2  r1 )  t r1 M1SAT B1 where r2 is the radial distance to the tube wall (r2=5.5mm) The percentage of beads that should be captured was calculated, dependent on the assumption of a well-mixed suspension (i.e., no spatial variation in bead concentration). The percentage of retained beads was calculated and compared to the percentage of retained beads predicted by the mathematical model.

Since the time for retention is a function of radial position, the number of beads captured is ultimately a function of their initial starting position. By calculating the volume of solution that should be emptied of beads in the given time, the number of retained beads was predicted. For example, for a residence time of 35s, the model predicts that all beads located from r=2mm to r=5.5mm (tube wall) will be captured. In order to calculate the number of beads retained, the volume of the cylinder from r=0mm to r=2mm was calculated and subtracted from the total volume. This gives the volume of the ring that should be emptied of beads. Given the initial concentration of beads, the number of beads retained and non-retained is calculated and compared to the actual bead counts performed experimentally. Two concentrations were tested to ensure there was no concentration effect on retention capability: 7x106 beads/mL (data set 1-2) and 20x106 beads/mL (data set 3-4). No effect was seen, as similar retention percentages were measured between the two concentrations.

Figure 4.6 and Table 4.2 show the results of these calculations. Figure 4.6 illustrates the percent of beads captured at each timepoint for the model predicted values as well as the

136 experimentally obtained values for the standardized beads. Table 4.2 compares the retention time, radial position, and percent of beads captured that were predicted by the model and measured in Runs 1-4. Generally, the retained percentages are similar to those predicted by the model. The experimentally measured values were compared to the model generated values using a two-sided independent one-sample t-test (significance with p < 0.05). Significant differences were observed at 5s and 45s. Variability is partially attributed to experimenter error in the timing of pipetting out the solution.

Table 2: Calculations for Retention of Standardized Micromer Beads Variable Value Units Description

rWALL 0.55 cm radius of tube/distance to wall 3 Vtotal 1 mL or cm total volume in tube h 1.05 cm height of solution in tube  3.14159265 pi 6 Cbeads R1-2 20 x10 /mL concentration of beads - Run 1&2 x106/mL Cbeads R3-4 7 concentration of beads - Run 3&4

Retention Position 106 Beads Percent of Beads Captured

Time r (center) VNegative VCapture Captured Predicted Run 1 Run 2 Run 3 Run 4 Average Standard s cm cm3 cm3 x106 % % % % % % Deviation 60 0.01 0.0003 0.9997 19.99 100.0% 97.3% 93.8% - - 95.5% 2.5% 45 0.1 0.0331 0.9669 19.34 96.7% 91.2% 87.9% 82.5% 81.2% 85.7% 4.7% 35 0.2 0.1322 0.8678 17.36 86.8% 86.2% 79.0% 80.0% 71.4% 79.2% 6.1% 25 0.3 0.2975 0.7025 14.05 70.2% 75.4% 62.6% 67.4% 57.2% 65.7% 7.7% 15 0.4 0.5289 0.4711 9.42 47.1% 52.1% 48.3% 44.0% 48.0% 48.1% 3.3% 10 0.45 0.6694 0.3306 6.61 33.1% 44.5% 39.8% 33.6% 25.7% 35.9% 8.1% 5 0.5 0.8264 0.1736 3.47 17.4% 28.4% 23.8% 21.4% 20.1% 23.5% 3.7%

Table 4.2: Calculation for retention of standardized beads in the EasySepTM magnet: theoretical vs actual values.

137 Percentage of Micromer Beads Retained over Time: Measured vs Predicted Values

100% * 90%

80%

70%

60%

50%

40% Measured Run 1 30% * Measured Run 2 Predicted

20% Measured Average PercentofBeads Captured Measured Run 4 10% Measured Run 3

0% 0 5 10 15 20 25 30 35 40 45 50 55 60 Time (s)

Figure 4.6: Percent of beads captured in the EasySep magnet during varying residence times (range 5s-60s). The curve predicted by the model is shown (blue line), as well as the experimentally measured data points and the average values (purple line). Asterisks indicate significant deviations (p < 0.05) in the predicted and measured values. By 60 seconds over 95% of the beads were retained in the magnet.

4.4.2 Cell Tracking Velocimetry (CTV)

The protocol and experimental setup of CTV has been described in detail elsewhere.10-12

The output of the system is the velocity of the labeled cells and their magnetophoretic mobility.

The histogram of frequency (normalized to the total number of cells tracked) versus magnetophoresis of the magnetized cells is given in figure 4.7. The sedimentation rate can also be measured with CTV and was calculated as 2.1x10-3 mm/s. For comparison, images of the magnetically induced velocity and sedimentation (figure 4.8) as captured

138 by the CTV system are presented. Note that the timescale for the images is different, since the faster velocity of the magnetically labeled cells required a greater number of frames per second. For the labeled cells in the magnetic field, 20 frames per second were required, and during sedimentation was 1 frame/second. In order to directly compare the images, the magnetically induced velocity should be multiplied by a factor of 20. As demonstrated here and in Appendix B, the sedimentation force is orders of magnitude smaller than the magnetic force, which justifies neglecting its contribution in the initial force equations.

Magnetophoretic mobility of magnetized cells 30 Median: 0.000444

25 Mean: 0.00104

15 Frequency 10

0 -0.0004 0.0006 0.0016 0.0026 0.0036 0.0046 0.0056 0.0066 0.0076 0.0086 0.0096 Magnetophoretic mobility m (mm3/T*A*s)

Figure 4.7: Magnetophoretic mobility of the magnetized cells as measured by the CTV system. The mean and median values, used for calculating the retention time, are noted.

139 Figure 4.8: Images of the magnetically induced velocity (A) and sedimentation (B) as captured by the CTV system. Note that the timescale for the images is different, since the faster velocity of the magnetically labeled cells required a greater number of frames per second (F/s). In order to directly compare the images, the cells in image A would have moved 20 times farther in the time span of image B.

The mean magnetophoretic mobility of the labeled calls was 1.04x10-12 m3/(T*A*s). This value was used in the model to predict the velocity of the cells in the EasySep magnet as well as the retention time of the labeled cells as a function of their radial position r. The calculated magnetization of the labeled cells and their velocity in the EasySep magnet

(presented on separate y-axes) as a function of position is presented in Figure 4.9. The calculated time required for a labeled cell to reach the tube wall and be retained given its initial radial position is presented in Figure 4.10. The retention time suggested by the manufacturer and used in the current cell separation protocol is 5 minutes. As demonstrated by this graph, that is more than sufficient for separation and retention of labeled cells, and in fact can be decreased to help minimize the overall preparation time required to obtain the final population of labeled cells.

140 The median magnetophoretic mobility of the labeled calls was 4.44x10-13 m3/(T*A*s).

The model predicts a maximum retention time of 119 seconds to retain labeled cells within the magnet.

Radial Distance versus Magnetization and Velocity

350000 1.60E-04

300000 1.40E-04

1.20E-04 250000 1.00E-04 200000 8.00E-05 150000

6.00E-05 Velocity (m/s) 100000 M 4.00E-05 Magnetization M (A/m) v 50000 2.00E-05

0 0.00E+00 0.000001 0.001 0.002 0.003 0.004 0.005 Center of tube Radial Distance r (m) Wall of tube

Figure 4.9: Magnetization and velocity of the magnetized cell as a function of radial position r.

141 Predicted Time Required for Magnetized Cell to be Retained in Magnet

140

120 mean median 100

40 Time to reach Time(s) wall to 20

0 1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 Center of tube Radial distance r (m) Wall of tube

Figure 4.10: Predicted time for the magnetized cells to be retained in the magnet as calculated by the model equations using both the mean and median magnetophoretic mobilities. Note r=0m indicates the center of the tube and r=0.0055m is the tube wall. Essentially all cells are retained by 60 seconds (mean) or 120 seconds (median).

4.5 Discussion:

A model was created to predict the required retention time of the magnetized cells in the

EasySep magnet and was tested with standardized Micromer beads. The Micromer beads showed fairly good agreement with the predicted model values, with two time points differing in the predicted versus measured values at 5s and 45s. Most of the variability in the measured values was attributed to experimenter error in the accurate and quick removal of the aqueous buffer solution at the required time point.

The derivation of the model required several assumptions. The displacement due to thermal motion and sedimentation was much smaller than that of the magnetically

142 induced motion, as shown in Appendix B, providing the sufficient difference in magnitude required to neglect these effects. In addition, Appendix A demonstrates that the magnetized cell reaches equilibrium essentially instantaneously, allowing the inertial term to be neglected. Now the cell mobility is dependent on the magnetic force and opposing drag force. This enables the model to be simplified by equating the magnetic force and drag force and using the CTV magnetophoretic mobility measurement to solve for the retention time required to retain magnetized cells against the tube wall.

Using the mean magnetophoretic mobility, the model calculations of the required residence time of the magnetized cells in the EasySep magnet show that the manufacturer-recommended time of 5 minutes was not only sufficient, but 5 times longer than necessary, providing room to reduce the overall time required for separation of the positive fraction of cells. When the median magnetophoretic mobility was used in the model, a maximum retention time required to isolate magnetized cells against the wall of the tube was 120 seconds, still 60% shorter than the manufacturer recommended time.

Using a simple CTV measurement to provide the magnetophoretic mobility of the cells, this model has utility in calculating the retention time required for magnetized cells. The protocol for labeling can be varied by methods including: reducing incubation times, decreasing antibody label concentrations, or exchanging magnetic beads tags entirely, and the model will report the new retention time required for the magnetized cells.

Alternately, cell suspensions can be labeled for new or different markers and the model will provide the necessary retention time in the magnet.

143 4.6 Appendi

4.6.1 Appendix A: Justification: Neglecting the Inertial Term

For small particles in liquid viscous media, the inertial term can be neglected so that ma =

0.8 The basis for neglecting this term can be seen by comparing the magnitudes of terms:

 F  FM  FD  ma

From the derivation section, FM and FD forces are:

FM  6Rv  ma

dv d 2r where a   . Substituting in: dt dt 2

dv F  6Rv  m M dt

Making dimensionless by defining the arbitrary scale factors to and vo,

t v   ,  to vo

d vo FM  6Rv o  m d to

Rearranging:

FM to 6R d  to  mvo m d

Now the terms dependent on time (dr/dt, d/dt) can be evaluated in terms of their coefficients. After algebraic manipulation, the acceleration coefficient (d/dt) is on the order of magnitude of one. If acceleration is a meaningful term, the coefficient of the velocity term will also be on the order of magnitude of one. Calculating the coefficient in front of the velocity term (Table A.1):

144 6R t  29163t (s1 ) m o o

-5 If the time scale to was on the order of 3.4x10 s, then the acceleration and velocity terms would be of the same order of magnitude. Since the time scale in this model is in seconds, the velocity coefficient is 4 orders of magnitude larger than the acceleration coefficient

(1). Therefore velocity term dominates the acceleration term and acceleration can be neglected. Practically speaking, this indicates that the velocity of the magnetized cell is reached almost instantaneously, and the magnetized cell‟s magnetically induced motion through the aqueous media is opposed only by drag.

Table A.1: Order of Magnitude Analysis

dv FM to 6R d FM  6Rv  m  to  dt mvo m d Symbol Value Units Description Source

FM - N Force due to magnetic field a dv/dt m/s2 Acceleration Derivative of velocity with respect to time R 4.85E-06 m Cell radius Average radius for BMMNCs (ViCell)  9.57E-04 kg/(m*s) Viscosity of buffer Measurement in Zborowski Lab  3.141592654 - Pi Constant 13 mCell 3.00E-12 kg Typical mass of a cell Reference Park K. et. al.

vo - m/s Velocity constant Dimensionless Analysis 6R 29163 s-1 Coefficient of velocity term Dimensionless Analysis m to 3.43E-05 s Time constant Dimensionless Analysis

Table A.1: Order of magnitude analysis

4.6.2 Appendix B: Justification: Sedimentation << Magnetic Displacement

Deriving Sedimentation Equations:

From the force diagram in figure 6, buoyancy FB and drag forces FD oppose gravity FG in the y-direction.

FG = FD-FB

The drag force is defined again by the Stokes equation for creeping flow,

145 FD  6Rv while the buoyancy force is defined by:

4 F   gV  R3  g B fluid 3 fluid where  is the density of the fluid, g is acceleration due to gravity (9.8 m/s2), and V is the

4 volume. Assuming that the magnetized cell can be modeled as a sphere, then V  R3 3

The gravity force is defined by:

FG  mg  Vgcell

Substituting in for FD, FB, and FG,

Vgcell  6R Vg fluid

Solving for velocity v:

dz Vg (   ) 2 R 2 g    cell fluid  dt 6R 9 

This is the equation for the terminal velocity of settling. Using the parameter values in

Table B.1, velocity is calculated as 4.06x10-3 mm/s. With a residence time in the EasySep magnet of 5 minutes, the displacement due to sedimentation is calculated at 1.22x10-3 m, or 1.22 mm.

Using the CTV apparatus, sedimentation of the cells was measured in the absence of the magnetic field. Velocity in this experiment for labeled cells was measured at 2.1x10-3 mm/s and for unlabeled cells was 2.0x10-3 mm/s. This gives a displacement due to sedimentation of 0.63 mm after 5 minutes. This is in good agreement with the calculated value presented above.

146

Table B.1: Sedimentation Displacement dz Vg (   ) 2 R 2 g    cell fluid  dt 6R 9 

Symbol Value Units Description Source V 4.78E-16 m3 Volume of cell Assumed spherical: 4/3**r3 3 fluid 984.2 kg/m Density of buffer Measurement in Zborowski Lab 3 14 cell 1060 kg/m Density of cell Reference Milne D. et. al.  76 kg/m3 Change in density Calculated R 4.85E-06 m Cell radius Average radius for BMMNCs (Vicell) g 9.81 m/s2 Gravity acceleration Constant  9.57E-04 kg/(m*s) Viscosity of buffer Measurement in Zborowski Lab t 300 s Residence time in magnet Recommended by Stem Cell Technologies v 4.06E-06 m/s Terminal velocity of settling Calculated xg 1.22E-03 m Displacement due to sedimentation Calculated

Table B.1 Calculations for sedimentation displacement.

Generally, for paramagnetic particles, the displacement due to the magnetic field xM is:

xM  mSM t where

V m  and 6R

dB S  H m dx where ∆ is the change in magnetic susceptibility between the buffer and magnetite, H is the magnetic field strength, and dB/dx is the magnetic field gradient.7

Superparamagnetic particles have a magnetic susceptibility much greater than that of the

-6 buffer (Magnetite≈90 and Water= -9.04x10 ) and so the effective susceptibility approximates that of magnetite alone. The susceptibility of the particle Magnetite is also equal to M/H.7 Plugging in:

MV MV dB m  and x  t 6RH M 6R dx where M is the volume magnetization. Now xM is:

147 MV dB x  t  mS t M 6R dx M

Using CTV and the measured magnetophoretic mobility to calculate the average total volume of magnetic carrier per cell (V):

1 mS 6R dB V  M M  dx 

With an average V established, the EasySep parameters can be plugged in to find the displacement in the EasySep magnet due to the presence of the magnetic field:

MV dB x  t M 6R dx

The calculated velocity in the EasySep magnet is 0.14 mm/s and the displacement xM is

46mm given a residence time of 5 minutes (Table B.2). This value is much larger than the sedimentation displacement (0.63 mm), therefore the assumption for neglecting sedimentation is justified.

Table B.2 Magnetic Displacement

1 MV dB mSM 6R dB xM  t  mSM t V  6R dx M  dx 

Symbol Value Units Description Source 3 m1 1.04E-12 m /(T*A*s) CTV magnetophoretic mobility Measured in CTV experiment 2 SM1 1.40E+08 T*A/m Magnetophoretic driving force Constant in CTV system t 300 s Residence time in magnet Recommended by Stem Cell Technologies XMCTV 4.36E-02 m Displacement (magnetic) in CTV Calculated M1SAT 342435 A/m Magnetization (saturated) in CTV Calculated from curve with B=140.3 T/m (Figure 8) dB/dx 140.3 kg/(m*s2*A) CTV gradient of B Constant in CTV system R 4.85E-06 m Cell radius Average radius for BMMNCs (ViCell)  9.57E-04 kg/(m*s) Viscosity of buffer Measurement in Zborowski Lab  3.141592654 Pi Constant V 2.64E-19 m3 Total volume of magnetite per cell Calculated dB/dr 148.12 kg/(m*s2*A) EasySep gradient of B Experiment - slope of trendline (Figure 5) XMEasySep 4.60E-02 m Displacement (magnetic) in EasySep Calculated (using MSAT)

Table B.2 Calculations for magnetic displacement.

148 Finally, the last effect on the cells is random thermal motion. Proving that the magnetic displacement xM >> {x} thermal displacement is required to prove the motion of the cells is magnetically induced.

xM >> {x}

From the Einstein formula and the Stokes-Einstein equation:

x 2Dt Einstein formula

kT D  Stokes-Einstein Equation 6R where {x} is the root-mean-square displacement due to thermal Brownian motion, D is the diffusion coefficient, k is the Boltzmann constant (1.381x10-23 J/K), and T is the absolute temperature. Table B.3 illustrates the calculations. The displacement due to thermal motion was calculated at 5.3x10-3 mm, which is 4 orders of magnitude less than the calculated displacement of magnetized cells (46mm). Displacement in the magnetic field can be contributed with confidence to magnetically induced displacement and not random thermal motion.

Table B.3: Random Thermal Displacement

kT x 2Dt D  6R Symbol Value Units Description Source k 1.38E-23 J/K Boltzmanns constant Constant T 297 K Absolute temperature Room temperature R 4.85E-06 m Cell radius Average radius for BMMNCs (ViCell)  9.57E-04 kg/(m*s) Viscosity of buffer Measurement in Zborowski Lab  3.141592654 Pi Constant D 4.69E-14 m2/s Diffusion coefficient Calculated t 300 s Residence time in magnet Recommended by Stem Cell Technologies {x} 5.30E-06 m Displacement due to thermal effects Calculated

Table B.3 Calculations for random thermal displacement.

149 4.7 References

1. Muschler, G.F., Boehm, C. & Easley, K. Aspiration to obtain osteoblast progenitor

cells from human bone marrow: the influence of aspiration volume. Journal of Bone

and Joint Surgery (American) 79, 1699-709 (1997).

2. Majors, A.K., Boehm, C.A., Nitto, H., Midura, R.J. & Muschler, G.F.

Characterization of human bone marrow stromal cells with respect to osteoblastic

differentiation. Journal of Orthopaedic Research 15, 546-57 (1997).

3. Copelan E. Hematopoietic Stem Cell Translantation. The New England Journal of

Medicine 354, 1813-1826 (2006)

4. Marquez-Curtis, L., Turner, A.R., Larratt, L.M., Letcher, B., Lee, S.F., &

Janowska-Wieczorek A CD34+ cell responsiveness to stromal cell–derived factor-

1a underlies rate of engraftment after peripheral blood stem cell transplantation

Transplantation & Cellular Engineering 49, 161-169 (2009)

5. Hatch, G.P. & Stelter R.E. Magnetic design considerations for devices and particles

used for high-gradient magnetic separation (HGMS) systems Journal of Magnetism

and Magnetic Materials 225, 262-276 (2001)

6. Fishbane P., Gasiorowics S. & Thornton S. Physics for Scientists and Engineers. 2nd

Edition. Prentice Hall 1993 863-879

7. Zborowski M. & Chalmers J.J., Magnetic Cell Separation: A volume in laboratory

techniques in biochemistry and molecular biology. Elsevier 2007

8. Warnke K.C. Finite-element modeling of the separation of magnetic microparticles

in fluid IEEE Transactions on Magnetics 39(3), 1771-1777 (2003)

150 9. Yamaura, M., Camilo, R.L., Sampaio, L.C., Nakamura, M.A. & Toma H.E.

Preparation and characterization of (3-aminopropyl) triethoxysilane-coated

magnetite nanoparticles Journal of Magnetism & Magnetic Materials 279 210–217

(2004)

10. Chalmers J.J., Zhao Y., Nakamura M. et al. An instrument to determine the

magnetophoretic mobility of labeled biological cells and paramagnetic particles.

Journal of Magnetism & Magnetic Materials 194, 231–41 (1999)

11. Chalmers J.J., Haam S., Zhao Y. et al. Quantification of cellular properties from

external fields and resulting induced velocity: magnetic susceptibility.

Biotechnology & Bioengineering 64(5), 519–26 (1999)

12. Moore L.R., Zborowski, M., Nakamura, M., McCloskey, K., Gura, S., Zuberi, M.,

Margel, S., Chalmers, J.J. The use of magnetite-doped polymeric microspheres in

calibrating cell tracking velocimetry. Journal of Biochemical & Biophysical

Methods 44, 115 –130 (2000)

13. Park, K., Jang, J., Irimia, D., Sturgis, J., Lee, J., Robinson, J.P., Toner M. & Bashir

R. „Living cantilever arrays‟ for characterization of mass of single live cells in

fluids The Royal Society of Chemistry: Lab on a Chip 8, 1034–1041 (2008)

14. Milne D.B., Ralston N, Waliwork J.C. Zinc Content of Cellular Components of

Blood: Methods for Cell Separation and Analysis Evaluated Clinical Chemistry

31(1), 65-69 (1985)

151 Chapter 5:

Development of protocol for magnetic separation based on hyaluronan for enrichment of connective tissue progenitors.

5. Chapter Introduction

In order to enable in vivo testing of the performance of MS-processed cells, a protocol must be developed to process large volumes of freshly aspirated bone marrow. The small scale EasySepTM magnet can process 2.5 mL of bone marrow in a single procedure, however, up to 48 mL of bone marrow is routinely aspirated and loaded into bone grafts in the clinical setting of large bone defects. In order to test the performance of HA+++ cells in a biologically relevant large animal model, up to 48 mL of marrow must be magnetically separated.

This chapter details experiments undertaken to develop a protocol and magnetic separation system capable of processing the large volumes of marrow required before implementation of in vivo testing in the canine femoral multidefect model (CFMD).

Experiments include:

1. Evaluation of LymphoprepTM preprocessing protocols to eliminate red blood cells from fresh marrow aspirates

2. Evaluation of a simplified single pass magnetic protocol against the previously established three pass protocol to minimize residence time in the magnet

152 3. Effect of primary and secondary label concentration on the retention of magnetized cells and CTP prevalence in the HA-positive fractions

4. Effect of preincubation of labeling reagents to minimize the number of labeling steps and incubation time required to magnetically label the cells before separation in the magnet

Scale up of magnetic separation was achieved using the hexapole magnetic system

(HMS), developed in the Zborowski laboratory at the Cleveland Clinic. The HMS provides a portable permanent magnet with the capacity to separate 50 mL of marrow in a single procedure. This magnet features a hexapole design and generates a maximum magnetic field of 0.64 teslas, while the EasySepTM magnet, used in Aim 1, is a quadrupole magnet that generates a maximum field of 0.74 teslas. The custom channel used in the HMS magnet features a central rod that blocks the center of the tube where the magnetic field strength is weakest, which enables a shorter retention time in the magnet since cells are excluded from the weakest part of the magnetic field. These magnets are similar in field strength; however, the capacity of the HMS is 20 times that of the EasySepTM magnet.

Experiments executed to test the performance of the HMS magnet using human and canine marrow samples were performed in the Zborowski lab by Pow Joshi and are detailed in a manuscript in preparation entitled: Scale-up separation of connective tissue progenitors of osteogenic potential from bone marrow aspirates using a novel design of

153 magnetic separation device and hyaluronan as an osteogenic progenitor marker. The final protocol used for experimentation in the CFMD is detailed in Chapter 6.

5.1 Evaluation of a method for RBC removal:

LymphoprepTM Separation

5.1.1 Introduction

The act of removing red blood cells (RBCs) introduces preprocessing steps that alter the composition of a fresh bone marrow aspirate, and often results in the loss of select cell populations which can include CTPs. In Aim 1, a repeated “buffy coat” density gradient separation was used to eliminate the majority of RBCs before labeling for hyaluronan

(HA) and subsequent magnetic separation processing. After centrifugation of marrow aspirates, the higher density elements of the marrow including RBCs pellet at the bottom of the tube. The plasma layer moves to the top of the tube, and at the interface of these two layers lies the buffy coat, so named due to its “buff” coloring (Figure 5.1).

Because the buffy layer abuts the RBC pellet, a clean aspiration of the entire buffy layer without including RBCs is challenging. The very top of the RBC layer is usually aspirated along with the buffy coat to ensure the harvest of all the nucleated cells of interest, including the CTPs. After a single buffy coat preparation, 98.4 ± 2.1% of the

CTPs were recovered in the buffy coat fraction. Only 0.51 ± 1.5% of the CTPs were found in the RBC fraction and 1.1 ± 1.6% of the CTPs partitioned to the plasma fraction

154 (N=28, unpublished data from Muschler lab), indicating that the overwhelming majority of CTPs were harvested in the buffy coat fraction. However, even with a repeated buffy coat separation, RBCs remain in the cell suspension.

Buffy Coat LymphoprepTM Separation Separation

Fat

Plasma Plasma

Buffy Cell coat LymphoprepTM RBC RBC

Figure 5.1: Illustration of buffy coat density separation and LymphoprepTM density gradient separation. In the buffy coat separation, the buffy coat layer contains the cells, and this layer is harvested and is used for further experimentation. In the LymphoprepTM separation, the Lyphoprep solution settles between the RBC layer and the cell layer, ensuring cleaner separation of cells from RBCs. The cell layer is harvested and used for further experimentation.

LymphoprepTM density gradient separation is often used to eliminate RBCs in peripheral blood draws and marrow aspirates, and provides a cleaner separation of cells from RBCs due to the density (1.077 g/mL) of the medium. In 1968, Dr. Arne Bøyum first developed a method for the isolation of mononuclear cells from human blood.1 Mononuclear cells, including monocytes and lymphocytes, have a lower buoyant density (< 1.077 g/ml) than the RBCs and polymorphonuclear leukocytes, which include eosinophils, neutrophils,

155 and basophils. Mononuclear cells are retained at the LymphoprepTM medium interface after centrifugation, while the erythrocytes and the PMNs sediment through the medium to the bottom of the tube (Figure 5.1).

In this study, density gradient separation based on LymphoprepTM was evaluated for use as a possible preprocessing method to remove RBCs from marrow aspirates before magnetic separation labeling and processing. Three protocols for LymphoprepTM separation were evaluated and compared using freshly aspirated human bone marrow aspirates. We hypothesize that LymphoprepTM processing will result in fractioning of

CTPs outside of the cell layer after density gradient separation.

5.1.2 Methods

5.1.2.1 Cell Source

Human bone marrow was aspirated from the iliac crest of 4 patients in 2 mL aliquots according to an institutional review board-approved protocol with informed consent.

5.1.2.2 Description of LymphoprepTM Protocols

Density gradient separation using LymphoprepTM (NC9223598, Axis-Shield, Oslo,

Norway) was evaluated and compared using three starting samples derived from freshly aspirated human bone marrow: whole marrow, diluted marrow, and a buffy coat preparation from whole marrow.

156 For whole marrow separation, 6 ml of freshly aspirated marrow was layered over 3 ml of

LymphoprepTM in a 15 ml Falcon tube. For diluted marrow separation, 3 ml of whole marrow was diluted with 3 ml of sterile saline and mixed. This 6 ml mixture of diluted marrow was layered over 3 ml of LymphoprepTM. Finally, for the buffy coated marrow separation, a buffy coat was prepared by density gradient separation (400 x g for 10 minutes) and the buffy cell layer was removed. Six milliliters of buffy coated marrow was layered over 3 mL of LymphoprepTM solution.

All 3 samples (whole, diluted, and buffy coated) were spun in an Allegra 6 Beckman

Coulter centrifuge (Beckman Coulter #366802, Brea, CA) at 800xg (1871 RPM) for 30 minutes at room temperature. Four fractions were collected after centrifugation: the top fatty layer, the cell layer, the red blood cell layer, and the remaining middle layer, consisting of a pooled solution of plasma, located between the top and cell layers, and

LymphoprepTM solution, located between the RBC and cell layers (Figure 5.2). Each fraction was diluted with 10 ml of media and spun for 10 minutes at 229 x g to remove the remaining LymphoprepTM solution.

157 Middle Layer

Fat Layer Plasma + Cell Layer LymphoprepTM RBC Layer

Figure 5.2: Illustration of the 4 fractions harvested after LymphoprepTM separation and assayed for cell and CTPs.

5.1.2.3 Cell Culture Conditions

Samples from each of these 4 fractions were stained with trypan blue (#T8154 Sigma

Aldrich, St. Louis, MO) for viability, placed in 0.3% acetic acid to lyse RBCs, and counted with a hemacytometer. All fractions, as well as the whole marrow and buffy coated marrow samples, were cultured under osteogenic medium conditions, consisting of -MEM with 10% fetal bovine serum, 1 unit/mL penicillin, 0.1 mg/mL streptomycin,

10-8 M dexamethasone, and 50 g/mL ascorbate. Each fraction was plated at a density of

500,000 cells per Lab-Tek chamber (4.2 cm2) (#177380 Nalge Nunc International,

o Rochester, NY), and cultured at 37 C at 5% CO2 with medium changes on Days 2 and 3.

Day 6 cultures were fixed with 1:1 acetone:methanol for 10 minutes and stained for nuclei (DAPI) and osteoblastic activity (alkaline phosphatase).

158 5.1.2.4 ColonyzeTM Image Acquisition and Analysis

Lab-Tek culture wells were scanned using a Spot RTSE 9.0 Monochrome-6 12 bit digital camera (Diagnostic Instruments Inc., Sterling Heights, MI) mounted on a Leica

DMRXA2 motorized microscope (Leica, Allendale, NJ) controlled by Metamorph (v6.3) imaging software (Universal Imaging Corporation, Downington, PA). The camera/microscope complex acquired 540 individual images of each culture well. Each image was collected at 340x256 8-bit gray level using a 10x objective (pixel size = 3.27

µm). A blank image was taken and used to background correct each individual image.

The individual images were then montaged to create a single image. A region of interest was applied to exclude debris around the edges of the chamber. Lint debris, apoptotic debris and glass aberrations were excluded by size and shape segmentation. The

ColonyzeTM image analysis software, used for quantitative characterization of CTPs, identifies colonies containing eight or more cells in a cluster. Each montaged image was reviewed by a single user. In the cases where debris, skipped colonies, and/or incomplete colony detection were found, the algorithm colony assessment was manually edited to correct these errors.

5.1.2.5 Statistical Analysis

All statistics were performed using an ANOVA. Post-hoc analysis was performed with multiple comparisons using Tukey‟s test with p<0.05.

5.1.3 Results

5.1.3.1 Cell Counts and Yield after LymphoprepTM Processing

159 The total cell yield after LymphoprepTM separation was 85.2 ± 26.7% for the buffy coated marrow (BCM), 68.0 ± 26.3% for the whole marrow (WM), and 103.5 ± 9.7% for the diluted marrow (DM). Cell counts were performed for nucleated cells in the fat, cell, red blood cell (RBC), and middle layers. Each count was divided by the starting count to determine the percent of cells partitioning to each fraction, and averaged over the 4 patients (Figure 5.3). Only 27.4 ± 8.8% (BCM), 18.6 ± 23.1% (WM), and 34.7 ± 32.6 %

(DM) of cells partitioned to the “cell” layer, and the majority of nucleated cells were in the RBC layer after centrifugation. There was no difference in cell partitioning between separation types (BCM, WM, DM). Most nucleated cells were recovered in the RBC fraction, which was significantly higher than the cell (p=0.012), fat (p<0.0001), and middle (mid) (p<0.0001) fractions. The cell fraction contained significantly more cells than the fat (p=0.025) and middle (p=0.024) fractions.

160 Figure 5.3: Partitioning of nucleated cells after LymphoprepTM preprocessing. Most cells were recovered in the RBC layer, followed by the Cell layer. Bars are standard error.

5.1.3.2 CTP Prevalence after LymphoprepTM Processing

The CTP prevalence for each layer is obtained after processing culture slides through the

ColonyzeTM system (Figure 5.4). CTP prevalence is defined as the number of CTPs per million cells plated. The CTP prevalence for each layer of the LymphoprepTM protocol was compared to each of the 3 bone marrow-derived starting samples. Compared to the buffy coat marrow, CTPs were enriched 1.24 ± 0.10 fold in the fat layer, and 1.13 ± 0.28 fold in the cell layer. Compared to the whole marrow, CTPs were enriched 2.34 ± 1.24 fold in the fat layer, and 1.70 ± 1.73 fold in the cell layer. Compared to diluted marrow,

CTPs were enriched 2.02 ± 1.20 fold in the fat layer and 0.93 ± 0.78 fold in the cell layer.

There was no significant difference between the BCM, WM, and DM separation types.

For all three samples, the fat layer had the highest CTP prevalence, followed closely by

161 the cell layer. The CTP prevalence in the fat layer is significantly higher than the mid layer and RBC layers (p=0.035 and p=0.006, respectively).

Figure 5.4: CTP prevalence, defined as the number of CTPs per million cells plated, in each fraction after LymphoprepTM separation. Bars are standard error.

5.1.3.3 Total number of CTPs in each layer after LymphoprepTM Processing The total number of CTPs can be calculated from the starting population. The number of total CTPs (CTPTOTAL) in the starting fraction can be obtained by multiplying the prevalence of the starting fraction (PS) by the starting number of cells (NS):

6 6 CTPTOTAL=PS (CTPs/10 cells) * NS (10 cells)

Similarly, the number of CTPs partitioning to each layer can be calculated by multiplying the layer prevalence by the cell count of that layer. Figure 5.5 illustrates the percent of total CTPs partitioning to each layer after lymphoprepTM separation. The total CTP yield after separation was 68.2 ± 10.9% for the BCM, 88.5 ± 24.2% for the WM, and 93.6 ±

162 54.4% for the DM. Only 29.5 ± 6.4% (BCM), 41.2 ± 51.8% (WM), and 48.2 ± 59.9 %

(DM) of cells partitioned to the “cell” layer. Between 30-40% of the CTPs were found in the RBC layer in all 3 types of separation.

Figure 5.5: CTP partitioning after LymphoprepTM separation. Between ~30-50% of the total CTPs partitioned to the cell layer after centrifugation (depending on the type of starting sample used). No significant difference was found between the BCM, WM, and DM sample types. Bars are standard error.

5.1.4 Conclusion

Typically, after LymphoprepTM separation, the concentrated “cell” fraction is removed and used for further experimentation. However, this data demonstrates that only 29.5 ±

6.4% (BCM), 41.2 ± 51.8% (WM), and 48.2 ± 59.9 % (DM) of CTPs partitioned to the

“cell” layer. Between 30-40% of CTPs were found in the RBC layer in all 3 types of separation.

163 LymphoprepTM is effective at removing RBCs but the loss of CTPs outside of the harvested “cell” layer is unacceptably high to be used as a preprocessing step before magnetic separation.

5.1.5 Discussion

Previous studies have demonstrated that density gradient separation using Ficoll-Paque™ or LymphoprepTM is an effective method of removing erythrocytes from bone marrow aspirates.2,3 LymphoprepTM and Ficoll-PaqueTM solutions share an almost identical composition.3 Ficoll-PaqueTM, marketed through GE Healthcare (Uppsala, Sweden) and

LymphoprepTM, sold through Axis-Shield (Oslo, Norway), both have a density gradient shelf of 1.077 ± 0.001 g/mL and a polysaccharide solution 5.7% weight per volume.

LymphoprepTM has a similar percent of sodium diatrizoate as Ficoll-PaqueTM

(LymphoprepTM: 9.1 w/v, Ficoll-PaqueTM: 9.0 w/v). One difference between the two products is that Ficoll-PaqueTM contains 0.0231 g of disodium calcium EDTA per 100mL while LymphoprepTM has no anticoagulant.3

Both Ficoll-PaqueTM and LymphoprepTM are equivalent in both RBC removal and cell recovery. Yeo et al. tested Ficoll-PaqueTM versus LymphoprepTM solutions and found no significant difference in recovery of any tested cell population or cell yield. Both were equally effective at removing RBCs. The number of RBCs remaining in the isolated cell fraction was 0.54 ± 0.38 x 108 using LymphoprepTM and 0.55 ± 0.34 x 108 using Ficoll-

PaqueTM.3 They concluded that there was no difference in the performance of Ficoll-

PaqueTM and LymphoprepTM solutions.

164

Both Ficoll-PaqueTM and LymphoprepTM solutions are efficient at RBC removal. Dragoni et al. compared 5 methods for preprocessing and RBC removal, including a Ficoll-

PaqueTM density gradient separation on a Cobe machine. Ficoll-PaqueTM on the Cobe density gradient separator removed the most RBCs compared to other available systems

(over 98%) and was efficient at PMN removal (over 98%) while maintaining a recovery of 89% of mononuclear cells and 80% of heamatopoietic CFU-GM.2

However, for applications in which the preservation of maximum numbers of CTPs are required, LymphoprepTM falls short as a preprocessing step. Our data demonstrates that this method is sub-optimal for isolation of CTPs, despite the use of different marrow preparations (whole marrow, buffy coat or diluted marrow), due to the combination of the number of CTPs lost in processing and the number of CTPs that remain in the erythrocyte fraction after LymphoprepTM separation. In line with these results, Kasten et al. tested

Ficoll-PaqueTM density gradient separation for recovery of mesenchymal stem cells

(MSC) and found that Ficoll-PaqueTM separation resulted in the loss of MSC. After Ficoll separation, the prevalence of MSC as assessed by CFU assay was significantly lower than unprocessed marrow. The MSC prevalence after Ficoll-PaqueTM was 50 CFU/mL bone marrow, while the MSC prevalence in unprocessed marrow was 156 CFU/mL bone marrow, indicating the loss of MSCs in layers other than the cell layer.4

165 5.2 Evaluation of a simplified single pass magnetic protocol

against the previously established three pass protocol to

minimize residence time in the magnet

5.2.1 Introduction

Using rapid intraoperative processing, a single surgical procedure is performed in which marrow is aspirated and loaded into a bone graft, then implanted into the bone defect site.

Similarly, in the canine femoral multidefect model, marrow is harvested and processed, and scaffolds loaded with cells are implanted in the canine in the same surgical procedure on the same day. Therefore, streamlining the magnetic separation protocol to reduce the time required for labeling and separation is desirable.

To simplify the protocol for HA separation, a single pass through the magnet was compared to the triple pass protocol used in Aim 1. If effective, a single pass would reduce the time required for magnetic separation from 15 minutes to 5 minutes.

We hypothesized that the single pass protocol will have a lower purity of HA-positive cells in the HA-positive fraction, but will still provide enrichment of CTPs in the HA- positive fraction.

5.2.2 Methods

5.2.2.1 Cell Source

166 Human bone marrow was aspirated from the iliac crest of 6 patients in 2 mL aliquots according to an institutional review board-approved protocol. Two sequential density gradient separations (“buffy coats”) were used to deplete red blood cells. Buffy coated marrow (BCM) cells were resuspended in buffer composed of phosphate buffered saline with 2% fetal bovine serum and 1 mM ethylenediaminetetraacetic acid (EDTA) at 200 million cells/mL.

5.2.2.2 Magnetic Separation Protocol using the EasySepTM system BCM cells (200x106/mL) were processed using the EasySepTM Magnetic Separation system (Stem Cell Technologies #18543) on the basis of HA expression using a biotinylated G1-link protein (hyaluronan binding protein, or HABP) (Calbiochem

#385911), which binds strongly and specifically to HA. Cells were sequentially labeled

TM with 200 L/mL of EasySep FC blocker, followed by 20 L/mL of biotinylated HABP at 0.5 mg/mL for 1 hr at room temperature. After removing excess HABP by centrifugation, the EasySepTM anti-biotin tetrameric antibody complex was added at 200

L/mL and incubated for 15 minutes, followed by EasySepTM magnetic nanobeads at 100

L/mL for a 10 minute incubation. After increasing the total volume to 2.5 mL using buffer solution, the cells were placed in the EasySepTM magnet for 5 minutes allowing magnetized cells to be pulled out of suspension and held against the walls of the container.

Two protocols were evaluated: the traditional 3-pass protocol and a simplified single pass protocol. The 3-pass protocol produced 3 fractions from the magnet: HA+++, HA+, and

HA- fractions. Cells retained in the magnet after 3 sequential separation steps were

167 labeled the HA+++ population. Cells that were not retained after each separation were sent back through the magnet. Any cells retained on a subsequent pass were identified as HA+.

Cells that were unbound on all passes through the magnet were defined as HA-. The simplified single pass protocol (1P) produced 2 fractions: HA+(1P) and HA-(1P). These fractions are produced after a single pass through the magnet.

5.2.2.3 Hematopoietic Progenitor Cell (HPC) Assay

To assay the partitioning of HPCs, a standard hematopoietic colony forming unit assay in methylcellulose was performed. Cells from each magnetically separated fraction

(100,000 nucleated cells in 100 L), as well as the unselected BCM, were combined with 2.5 mL of methylcellulose, 250 L of IMDM serum-free medium, and 150 L of a growth factor cocktail containing 50 ng/mL stem cell factor, 50 ng/mL granulocyte macrophage-colony stimulating factor (CSF), 50 ng/mL granulocyte-CSF, 10 ng/mL interleukin-3, 3 L/mL erythropoietin, 10 ng/L transforming growth factor FL-1, and 5 ng/L thrombopoietin. Each fraction was plated in duplicate (1 mL per well) and cultured

o in humidified 6 well plates cultured at 37 C at 5% CO2. Erythroid (CFU-E) and myeloid

(CFU-M) colonies were hand counted in each well after 14 days of culture. Prevalence, defined as the number of CFU-M or CFU-E per million cells plated, was calculated and standardized to the prevalence of the BCM.

5.2.2.4 Colony Forming Unit Assay for CTPs

Each fraction was counted with a hemacytometer. BCM, HA+++, HA+, HA- , HA+(1P), and HA-(1P) fractions were cultured in osteogenic medium consisting of -MEM with

168 10% fetal bovine serum, 1 unit/mL penicillin, 0.1 mg/mL streptomycin, 10-8 M dexamethasone, and 50 g/mL ascorbate (added on the day of use). Each fraction was plated at a density of 500,000 cells per LabTek chamber (4.2 cm2), and cultured at 37o C at 5% CO2 with medium changes on Days 2 and 3. Day 6 cultures were fixed with 1:1 acetone:methanol for 10 minutes and stained for nuclei (DAPI) and osteoblastic activity

(alkaline phosphatase, AP).5

5.2.2.5 ColonyzeTM Image Acquisition and Analysis

LabTek culture chambers were scanned using a Spot RTSE 9.0 Monochrome-6 12 bit digital camera (Diagnostic Instruments Inc.) mounted on a Leica DMRXA2 motorized microscope controlled by Metamorph (v6.3) imaging software and analyzed as previously published.5 The ColonyzeTM image analysis software was used to quantitatively identify colonies containing eight or more cells in a cluster using a standardized algorithm. CTP Prevalence was defined as the number of CTPs per million cells plated. Each individual colony was characterized based on cell count, colony area

(mm2), and alkaline phosphatase (AP) expression (AP area/cell number).

5.2.2.6 Flow Cytometry

#554061) to detect the biotinylated HABP and with Glycophorin A- fluorescein

169 isothiocyanate (FITC) (DakoCytomation, F0870) for detection of erthyroid cells, and then run on a Becton Dickinson LSRII flow cytometer and analyzed with FlowJo software. Secondary antibodies (SA-PE) or isotype controls (FITC anti-mouse IgG2b,

BD Pharmingen 553395) were used to determine gating for the experimental samples.

Gate application on glycophorin A-negative cells eliminated the numerous glycophorin

5.2.2.7 Statistical Analysis

Due to the known wide variation in CTP prevalence and performance between individual subjects, the median per patient per condition was obtained for each colony level metric and then standardized to the median of the unselected BCM as a control. These standardized values were log base 2 transformed to obtain a Gaussian distribution of the response. Means and 95% confidence intervals were calculated. Back transformation provided the geometric mean and the 95% confidence intervals for the geometric mean, providing the relative magnitude of change in the colony level metrics from the control

BCM. For analysis of cell counts and flow cytometry data, an ANOVA was used. Post- hoc analysis was performed with multiple comparisons using Tukey‟s test with p<0.05.

5.2.3 Results:

5.2.3.1 Cell and CTP yield and partitioning after magnetic separation

170 After magnetic separation, cells were counted in each fraction (HA+++, HA+, HA-,

HA+(1P), and HA-(1P)). The mean yield of total nucleated cells after the 3 pass magnetic separation (3P MS) was 71.2 ± 23.3%, indicating that 28.0% of the cells in the starting population were lost during processing. Of the nucleated cells recovered after 3P separation, 2.8 ± 2.1% of the cells were retained in the HA+++ population, 14.1 ± 8.5% were found in the HA+ population, and 83.0 ± 10.4% were HA-, which was significantly different from the HA+ and HA+++ fractions (p<0.0001).

The yield after one pass magnetic separation (1P MS) was 80.6 ± 25.6%, with a nucleated cell loss of 19.4%. The HA+(1P) contained 17.3 ± 6.3% and the HA-(1P) contained 82.7

± 6.3% of the nucleated cells recovered after 1P MS processing, which was significantly different (p<0.0001).

5.2.3.2 CTP Partitioning after magnetic separation

The range of CTP enrichment observed after magnetic separation is illustrated in Figure

5.6. The total number of CTPs (CTPTOTAL) was calculated by multiplying the prevalence of CTPs in the initial BCM sample (PS) by the total number of nucleated cells in the

+++ + - starting sample (NS). The total number of CTPs in the HA , HA , and HA populations was calculated using the same method. The loss of colony forming CTPs during

171 processing was greater than the loss of cells in general. These calculations suggest

Figure 5.6: ColonyzeTM images illustrating the partitioning of CTPs after magnetic separation processing. Representative images from 3 of the 6 patients are shown. The top row shows the CTP prevalence in the BCM for these patients, illustrating the typical variability seen between human donors. Images from the 3 pass separation are shown for

172 HA+++ and HA- fractions. The final rows are images from the 1 pass separation (HA+(1P) and HA-(1P)). The fold enrichment over the BCM is listed for HA+++ and HA+(1P) fractions in white.

that 64.4 ± 22.2% of the CTPs in the initial sample were lost during the 3P MS multiple processing steps (e.g. via adherence to surfaces and/or greater buoyancy following centrifugation), in contrast to a 28.0 ± 23.3% loss of nucleated cells. During the 1P MS,

72.0 ± 23.8% of the CTPs in the initial sample were lost during the 3P MS processing, in contrast to a 19.4 ± 25.6 % loss of nucleated cells (not significantly different).

Of the CTPs recovered after MS processing, the HA+++ fraction contained 26.6% of all

CTPs, but only 2.8% of nucleated cells. The HA- fraction contained the majority of nucleated cells (83.0%) and 63.8% of the CTPs. The HA- fraction contained significantly more CTPs than the HA+++ fraction (p= 0.0247) and the HA+ fraction (p=0.0285). The

HA+(1P) fraction contained 68.9% of the recovered CTPs while the HA-(1P) contained

31.1%.

5.2.3.3 Purity and Recovery assessed by flow cytometry

Flow cytometric analysis was used to assess the purity and recovery of the HA-positive cells after magnetic separation. Erythroid cells were labeled with a glycophorin-A antibody to provide a method for eliminating erythroid cells from analysis. The non- erythroid nucleated cell fraction was selected using a Glycophorin-A negative gate

(figure 5.7), and the HA-positive cells were subsequently quantified (figure 5.8).

173 Magnetic separation was effective in isolating cells that retained or expressed HA on their surface (n=8). The mean purity of the HA+++ fraction was 50.6 ± 25.1%. The mean purity of the HA+(1P) fraction was 26.6 ± 14.4%. A purity of 100% would indicate that all the cells isolated in the HA+++ fraction are positive for HA.

Percent of cells in each fraction that are non-erythroid

100%

90% p = 0.0055 80% p = 0.0004 p = 0.0023 70% p = 0.0003 60%

50%

40%

A negativenegative A A cellcell percent percent (%)(%) 30%

20% Glycophorin Glycophorin 10%

0% UnselectedBCM Marrow HA+(1P)1pass + HA1pass-(1P) - HAHA++ HAHA++++++ HAHA--

Figure 5.7: A gate on gylcophorin A negative cells was applied. The non-erythroid (glycophorin A negative) cell population is shown here for each MS-processed cell fraction. The HA+++ fraction contained the least amount of erythroid cells (<10%). Bars are standard error.

Magnetic separation was effective in recovering cells with HA on their surface. The mean recovery was 75.5 ± 12.9% in the HA+++ fraction alone. The sum of both HA- positive fractions (both HA+++ and HA+) contained 92.2 ± 9.5% of all cells presenting HA on their surface. A recovery of 100% would indicate that there are no HA-positive cells in fractions other than the HA-positive fraction(s). Therefore, less than 8% of all cells

174 found to stain positive for HA on their surface were recovered in the HA- fraction. In the

1 pass separation, the mean recovery was 84.1 ± 17.0%.

Bars are standard error Percent of non-erythroid cells expressing HA 100%

90%

80%

70% p < 0.0001 p = 0.0014 p < 0.0001 p < 0.0001 p < 0.0001

60%

positivepositivecells cells

positivecells 50%

- - -

40%

Percent of cells expressing HA expressing of cells Percent 30%

Percent ofof Percent Percent HA HA Percent of Percent HA 20% p = 0.0035 p = 0.0026 10%

0% Unselected Marrow 1pass + 1pass - HA+ HA+++ HA- BCM HA+(1P) HA-(1P) HA+ HA+++ HA-

Figure 5.8: The non-erythroid cells in each fraction are examined for hyaluronan expression to determine the purity of magnetic separation. The HA+++ cell fraction contained the most HA-positive cells (>50%). This was signifinicantly higher than every other MS-processed fraction. The HA+(1P) fraction contained a lower purity of 26%, which was still higher than the HA- fractions.

5.2.3.4 Partitioning of Colony Forming HPCs

HPCs were cultured for 14 days, then myeloid (CFU-M) and erythroid (CFU-E) colonies were counted to access the prevalence of myeloid and erythroid progenitors. Prevalence, defined as the number of myeloid or erythroid colonies per million cells plated, is presented standardized to the prevalence of the BCM. This corrects for the large variability in baseline prevalence between patients, and presents data in terms of enrichment or depletion when compared to the control (represented as the red line at 1).

175

Hematopoietic progenitor cells were negatively selected (depleted) based on HA expression. The relative prevalence of HPCs in the HA+++ fraction was 0.03 for CFU-M

(a 33 fold reduction from the BCM) and 0.11 for CFU-E (a 9 fold reduction from the

BCM) (Figures 5.9 and 5.10). Both myeloid and erythroid progenitors partition to the

HA- fractions, which show similar prevalence to the BCM.

The 1 Pass fractions follow the same trend. The HA-(1P) prevalence is not significantly different than the BCM. The CFU-M prevalence in the HA+(1P) fraction was 5-fold less than the BCM, and the CFU-E prevalence was 3.3-fold less than the BCM. Both were significantly lower compared to the BCM.

2.02.0

1.81.8

1.61.6

1.41.4

1.21.2

1.0 1.0

cells plated plated cells cells 6 6 0.80.8

0.60.6

per 10 10 per per 0.40.4

0.20.2

(Standardized control) control) BCM BCM the the to to (Standardized (Standardized Prevalence of Myeloid Progenitors of Myeloid Prevalence Prevalence of Myeloid Progenitors of Myeloid Prevalence 0.00.0 HA+++HAHA++++++ HAHA+HA++ HAHA-HA-- HAHA++HA+(1P) (1P) (1P) HAHA-HA-- (1P)(1P) (1P)

Figure 5.9: Prevalence of myeloid progenitors (CFU-M) in each fraction after magnetic separation. The HA-positive fractions are all significantly depleted in CFU-M compared to the BCM.

176 2.02.0

1.81.8

1.61.6 Progenitors Progenitors 1.41.4

1.21.2

1.0 1.0

cells plated plated cells cells

6 6 Erythroid Erythroid 0.80.8

0.60.6

per 10 10 per per 0.40.4

0.20.2

(Standardized control) control) BCM BCM the the to to (Standardized (Standardized Prevalence of of Prevalence Prevalence of of Prevalence 0.00.0 HA+++HAHA++++++ HA+HAHA+ + HAHA-HA-- HAHA++HA+(1P) (1P) (1P) HA-HAHA- (1P)-(1P) (1P)

Figure 5.10: Prevalence of erythroid progenitors (CFU-E) in each fraction after magnetic separation. The HA-positive fractions are all significantly depleted in CFU-E compared to the BCM.

5.2.3.5 Partitioning of CFU-M and CFU-E to MS-processed fractions

“Accounting” of CFU-M and CFU-E can also be performed, analagous to that calculated previously for CTPs. The total number of cells after magnetic separation was multiplied by the CFU-E or CFU-M prevalence to give the total number of CFU-E and CFU-M present after magnetic separation. The percent of total CFU-E and CFU-M found in each fraction was calculated. The percent of CFU-E lost from the BCM using the 3pass separation was 50.5%, while the 1 pass separation was 65.0%. The percent of CFU-M lost during 3Pass processing was 59.6% and 70.7% during 1Pass processing.

Since the HA- fractions contained the highest prevalence of HPCs as well as the majority of other nucleated cells, most of the total number of predicted HPCs are found in the HA-

177 fractions. The percent of CFU-E in the HA- fraction was 70.0 ± 23.6%, which was significantly higher than the HA+ (p = 0.0001) and HA+++ (p = 0.0001) fractions. The

HA-(1P) contained 40.3 ± 23.0% of the CFU-E after 1P MS, which was higher than the

HA+(1P) (p = 0.0297). The HA+(1P) fraction contained 4.8 ± 5.4% of CFU-E and the

HA+++ fraction contained the lowest percent of CFU-E (0.21± 0.14%).

The percent of CFU-M in the HA- fraction was 56.7 ± 29.0%, and in the HA-(1P) was

18.9 ± 8.7%. The HA+(1P) fraction contained 4.8 ± 5.4% of CFU-E and 1.7 ± 1.5% of

CFU-M. The HA+++ fraction contained the lowest percent of CFU-M (0.07 ± 0.07%). The percent of CFU-M in the 3P HA- fraction was significantly higher than all other MS- processed fractions.

5.2.3.6 Prevalence of Colony Forming CTPs

Using the 3P MS, HA+++ cells were significantly enriched in CTPs, with a mean enrichment of 2.7-fold over the BCM (Range of 95% C.I: 1.5-4.4), and a 7.5-fold enrichment over the HA- fraction. There is across-the-board enrichment in the HA+++ population when compared to both the BCM and the HA- fraction. The colony prevalence for each population is given in Figure 5.11, and is defined as the number of CTPs per million nucleated cells plated. The HA+ and HA- fractions were significantly depleted in progenitors (0.44-fold and 0.36-fold, respectively).

The HA+(1P) shows similar prevalence as the BCM control (1.2-fold, Range of 95% C.I:

0.7-1.8), while the HA-(1P) fraction is significantly depleted in CTPs (0.15-fold). Unlike

178 the HA+++ fraction, the HA+(1P) fraction was not significantly enriched in CTPs over the

BCM.

8 8

7 7 /million cells) /million /million cells) /million 6 6

5 5

CTPsCTPs

4 4

3 3

2 2

1 1

0 0

CTP Prevalence (# (# (# CTPCTPPrevalence Prevalence toto control)control) BCM BCM the the (Standardized (Standardized HA+++HAHA++++++ HAHA+HA++ HAHA-HA-- HAHA++(1P)HA+ (1P) (1P) HAHA-HA--(1P) (1P) (1P)

Figure 5.11: CTP prevalence in each fraction after MS processing. The HA+++ fraction shows CTP enrichment in every patient using the 3P MS. However, HA+(1P) population, produced by the 1P MS, is not significantly enriched in CTPs compared to the BCM. All other fractions (HA+, HA-, HA-(1P)) are significantly depleted in CTPs compared to the BCM.

5.2.3.7 Biological Performance of Colonies Derived form CTPs Isolated in the HA+++,

HA+, and HA- Fractions

5.2.3.7.1 Proliferation:

Proliferation, as measured by the number of cells per colony, is illustrated in Figure 5.12.

The progeny of CTPs in the HA+++ population demonstrated a significantly greater proliferation rate than CTPs in the BCM or in the HA- fraction. The colonies derived from CTPs in the HA+++ population contained a mean number of cells per colony that was 2.0-fold (Range of 95% C.I: 1.4-2.9) greater than the colonies derived from BCM.

179 The mean cells per colony in the HA+ population was not different than unselected marrow, while the HA- fraction had significantly fewer cells per colony than the BCM

(0.75-fold).

The progeny of CTPs in the HA+(1P) population have significantly more cells per colony than the BCM (1.3-fold, Range of 95% C.I: 1.1-1.6). The HA-(1P) fraction was significantly less proliferative than the BCM (0.66-fold).

3.53.5

3 3

2.52.5

2 2

1.51.5

1 1

0.50.5

Relative Cells per Colony per Cells Relative control) control) toto BCM BCM the the (Standardized (Standardized Relative Cells per Colony per Cells Relative 0 0 HA+++HAHA++++++ HAHA+HA++ HAHA-HA-- HAHA++HA+(1P) (1P) (1P) HAHA-HA-- (1P)(1P) (1P)

Figure 5.12: The number of cells per colony, an indicator of proliferation, is shown for each fraction after MS processing. The HA+++ and HA+(1P) fractions show significantly greater cells per colony than the BCM.

5.2.3.7.2 Alkaline Phosphatase Staining:

Alkaline phosphatase activity was examined to gauge the capacity for osteoblastic differentiation of the CTPs (Figure 5.13). The HA+++ fraction showed significantly increased alkaline phosphatase expression, indicative of osteoblastic differentiation.

180 Colonies derived from CTPs in the HA+++ fraction had 3.1-fold (Range of 95% C.I: 1.1-

6.8) greater AP expression (area per cell) than colonies derived from BCM, suggesting that the progeny of HA+++ CTPs can more quickly exhibit a CTP-O phenotype. The HA+ and HA- fractions expressed alkaline phosphatase levels similar to the BCM.

Alkaline phosphatase activity in the HA+(1P) fraction was not different than that of the

BCM. Progeny of the CTPs in the HA-(1P) fraction was significantly lower than the

BCM (0.63).

8 8

7 7

6 6

5 5

4 4

3 3

2 2

1 1

0 0

Relative AP Area/Cell NumberNumber Area/Cell Area/Cell AP AP Relative Relative toto control)control) BCM BCM the the (Standardized (Standardized HA+++HAHA++++++ HA+HAHA++ HA-HAHA-- HAHA++HA+(1P) (1P) (1P) HA-HAHA- (1P)-(1P) (1P)

Figure 5.13: Alkaline phosphatase expression, an indicator of osteoblastic differentiation, is quantitated as the AP area per cell number in each colony. AP expression was significantly increased after 6 days of culture in the HA+++ fraction, suggesting that the progeny from HA+++ CTPs can readily differentiate down the osteoblastic lineage.

5.2.4 Discussion

181 To simplify the protocol for HA separation, a single pass through the magnet was compared to the triple pass protocol. We hypothesized that the single pass protocol will have a lower purity of HA-positive cells in the HA-positive fraction, but will still provide enrichment of CTPs in the HA-positive fraction.

Magnetic separation isolated cells that retained or expressed HA on their surface. The percent of non-erythroid cells in the BCM expressing HA was 10.0%. The percent of cells labeled for HA and detected by flow cytometry was higher than the percent detected in the BCM. The purity of the HA+(1P) was lower than the HA+++ fraction (26.6% vs.

50.6%), which was expected, since repeated passes through the magnet would allow cells nonspecifically trapped and retained on the first pass to be eliminated in the HA- fraction on subsequent passes.

Magnetic separation using hyaluronan enriches the CTP population using the 3 Pass protocol. CTP were significantly enriched 2.7-fold in the HA+++ fraction over the BCM control and 7.5-fold over the HA- fraction. However, the CTP prevalence in the HA+(1P) fraction was not different than the BCM, suggesting that multiple passes through the magnet are required to further purify the HA-positive cells where the CTPs may be located.

The biological performance of the progeny of CTPs isolated from the HA+++ and

HA+(1P) differ in alkaline phosphatase activity but not in their proliferative capacity. In comparison to colonies formed by CTPs from the unselected BCM, progeny formed by

182 HA+++ CTPs and HA+(1P) CTPs are significantly more proliferative. Progeny of HA+++

CTPs showed increased osteoblastic differentiation while alkaline phosphatase expression in the HA+(1P) colonies was not different than the BCM.

The HA+(1P) fraction captures a larger percentage of CTPs in comparison to the HA+++ fraction, but represents a less pure population. Both of the HA-negative fractions (HA- and HA-(1P)) were depleted in CTPs and significantly less proliferative than the BMA control. The in vivo effect on bone formation from either of these populations is unknown.

It is hypothesized that increasing the prevalence and concentration of CTPs by positive selection will enable increased survival and performance in vivo due not only to the increased number of CTPs but also the reduction of non-osteogenic cells implanted.

It is clear that the HA+++ fraction selects a subset of all CTPs, and that the HA- population still contains a majority of CTPs (as well as the majority of all other bone marrow-derived cells). These highly proliferative HA+++ cells may still offer superior performance in an in vivo graft environment, due to the elimination of the majority of non-essential, non-osteogenic cells that compete with CTPs for the limited oxygen and nutrients available at the graft site. This overwhelming disparity in metabolic demand limits the depth at which CTPs can remain viable in the graft, and these competing, non- osteogenic cells contribute to persistent inflammation as pro-inflammatory cytokines and cell debris are released after cell death.

183 A large percentage of the CTPs are unaccounted for after processing, even though the majority of bone marrow mononuclear cells (>70% yield) are present and viable (per trypan blue viability testing). On average, 64% of CTPs are lost during the staining and magnetic separation process using the 3-pass protocol, and 72% using the 1 pass protocol.

It is unclear if some of the cells lose viability or the ability to attach to the culture slides after magnetization, but, given the robust colony formation of the adherent CTPs, this is doubtful. More likely, the inherent preference of CTPs to adhere to plastic surfaces is the cause, as the staining and magnetization procedure provides ample opportunities for the

CTPs to attach to the various equipment to which they are exposed. Future experimentation seeks to uncover the cause of this apparent CTP loss.

5.3 Effect of primary and secondary label concentration on the

retention of magnetized cells and CTP prevalence in the HA-

positive fractions

453.1 Introduction

Previously, selection with hyaluronan (HA) has demonstrated enrichment of CTPs in the

HA+++ cell fraction. As the amount of marrow to be processed increases, the amount of reagent required to label for HA increases concomitantly. For the in vivo testing in the canine femoral multidefect model, 48 mL of heparinized marrow must be processed, so the reagent load must also be scaled appropriately. Reduction of the amount of either, or both, the HA binding protein (primary label) and the magnetic tag (magnetic nanobead and EasySep antibody, secondary label) would reduce the cost of each in vitro and in vivo

184 scaled up experiment. However, effective labeling of HA-presenting cells must be preserved.

This study examines the effect of concentration of the HA binding protein and the magnetic tag on CTP selection and magnetic mobility. We hypothesize that as the amount of reagent is reduced the number of CTPs retained in the HA+ fraction will decrease.

5.3.2 Methods:

5.3.2.1 Cell Source

Bone marrow was aspirated from the humerus of 3 canine donors in 2 mL aliquots according to approved ACURO protocol #W81XWH-08-2-0034. Aspirates were collected in heparinized media. Two sequential density gradient separations (“buffy coats”) were used to deplete red blood cells. The resulting buffy coated marrow (BCM) cells were resuspended in phosphate buffered saline (PBS) with 2% fetal bovine serum

(FBS), 1 mM ethylenediaminetetraacetic acid (EDTA) and 2 units/mL heparin at 200 million cells/mL.

5.3.2.2 Magnetic separation and variation of amount of labeling reagents

BCM cells were processed using the EasySepTM Magnetic Separation system (Stem Cell

Technologies #18543) on the basis of HA expression using a biotinylated G1-link protein

(hyaluronan binding protein, or HABP) (Calbiochem #385911), which binds strongly and specifically to HA.

185 Table 5.1 illustrates the variation of staining parameters, where the normal amount of label (used in pervious experiments) is represented by 1. The dilution is varied by 1/3 or

1/9 of the original volume of reagent. The objective of this experiment is to optimize the amount of stain required to select CTPs from a fresh bone marrow aspirate. Cells were labeled with biotinylated HABP for one hour at room temperature (RT). After removing excess HABP by centrifugation, the EasySep anti-biotin tetrameric antibody complex was added and incubated for 15 minutes, followed by magnetic nanobeads for a 10 minute incubation. After increasing the total volume to 2.5 mL, the cells were placed in the EasySep magnet for 5 minutes allowing magnetized cells to be pulled out of suspension against the walls of the container. The nonmagnetized population HA- fraction was decanted. The tube was removed and the HA+ fraction was resuspended in buffer and removed. Three million cells were sent to CTV for measurement of their magnetophoretic mobility, while one million cells were reserved for cell culture.

TableTable 5.1:1: Systematic Systematic Variation of of Staining Staining Parameters Parameters

Tube 3 Tube 2 Tube 1 1 1/9 HC : 1 EC 1/3 HC : 1 EC 1 HC : 1 EC

Tube 4 1/3 1 HC : 1/3 EC

Tube 6 Tube 5

Concentration(EC) 1/9

1/9 HC : 1/9 EC 1 HC : 1/9 EC EasySep TAC and Bead EasySep

1/9 1/3 1

HABP Concentration (HC)

5.3.2.3 Cell Culture Conditions

186 HA+, HA- and the unselected BCM were cultured under osteogenic media conditions, consisting of -MEM with 10% fetal bovine serum, 1 unit/mL penicillin, 0.1 mg/mL streptomycin, 10-8 M dexamethasone, and 50 g/mL ascorbate (added on the day of use).

Each fraction was plated at a density of 250,000 cells per LabTek chamber (4.2 cm2), and

o cultured at 37 C at 5% CO2 with medium changes on Days 2 and 3. Day 6 cultures were fixed with 1:1 acetone:methanol for 10 minutes and stained for nuclei (DAPI) and osteoblastic activity (alkaline phosphatase, AP).

5.3.2.4 ColonyzeTM Image Acquisition and Analysis

Culture wells were imaged and analyzed with the ColonyzeTM image analysis software to obtain CTP prevalence. LabTek culture chambers were scanned using a Spot RTSE 9.0

Monochrome-6 12 bit digital camera (Diagnostic Instruments Inc.) mounted on a Leica

DMRXA2 motorized microscope controlled by Metamorph (v6.3) imaging software. The

ColonyzeTM image analysis software is used to quantitatively identify colonies containing eight or more cells in a cluster using a standardized algorithm.

5.3.3 Results

5.3.3.1 Partitioning of Cells after MS processing

After a single pass through the magnet, nucleated cells were counted. Figure 5.14 illustrates the percentage of cells partitioning to the HA+ fraction for each patient as well as the average (red circles). Reduction of the secondary label by 1/9 resulted in significant decreases in the number of cells captured in the HA+ fraction when compared to all variations on the primary label (1, 1/3, and 1/9). As the amount of secondary stain decreased, the number of cells partitioning to the positive fraction also decreased, since

187 some of the cells were no longer labeled with enough magnetic carrier to be retained in the magnetic field. There was no significant effect of decreasing the primary label alone.

Figure 2: Percent of cells captured in the HA+ fraction as a function of decreasing label 25%

20%

15%

10%

5% Percent of in Cells fraction HA+ Percent

0% 1:1 1/3:1 1/9:1 1:1 1:1/3 1:1/9 1/9:1/9 Dilution of Primary and Secondary Label Figure 5.14: Partitioning of cells to the HA+ fraction after magnetic separation as a function of decreasing primary and secondary label concentration. Decreasing the amount of secondary label reduced the number of cells partitioning to the HA+ fraction. For all significant changes shown, p < 0.05.

5.3.3.2 Magnetophoretic moility of cells after MS processing

Magnetophoretic mobility was measured by CTV. Figure 5.15 shows the mean magnetophoretic mobility of the positively labeled cells. While the variability between patients is high, there is no significant effect on the average mobility throughout the range of dilutions tested here.

188 Figure 3: Mean Magnetophoretic Mobility of Positively Labeled Cells 3.5

3.0

2.5

/T*A*s) /T*A*s) 2.0 3

mm 1.5 -4

(x10 1.0

MagnetophoreticMobility 0.5

0.0 1:1 1/3:1 1/9:1 1:1 1:1/3 1:1/9 1/9:1/9 Dilution of Primary and Secondary Label

Figure 5.15: Mean magnetophoretic mobility of the positively labeled cells in the HA+ fraction as a function of decreasing label concentration. No changes in mean mobility in the positive cells were observed.

5.3.3.3 CTP Prevalence

HA+ fractions from each dilution were cultured for 6 days and analyzed for colony formation using the Colonyze software system. Figure 5.16 illustrates the prevalence of

CTPs at each dilution. There was no significant effect on CTP prevalence for any of the dilutions tested. There was a trend of CTP prevalence to increase as the amount of stain decreased, suggesting that these cells are strongly magnetically labeled.

189 Figure 4: CTP Prevalence of the HA+ fraction

3.5

3.0

2.5

2.0

cells plated plated cells 6 1.5

1.0

CTPs per 10perCTPs 0.5 (Standardized to the 1:1stain) the to (Standardized 0.0 1:1 1/3:1 1/9:1 1:1 1:1/3 1:1/9 1/9:1/9 Dilution of Primary and Secondary Label

Figure 5.16: CTP Prevalence as a function of decreasing label concentration. No significant difference was detected in any of the samples tested, though there was a trend toward increased CTP prevalence as the secondary label concentration was reduced.

5.3.3.4 CTP Partitioning

Based on the initial prevalence of the bone marrow aspirate, the total number of CTPs present in the starting population can be predicted. After magnetic separation, the cell count for the HA+ fraction multiplied by the HA+ CTP prevalence gives the total number and percent of CTPs captured in each HA+ sample (Figure 5.17). There was no significant effect of label dilution on the percent of total CTPs partitioning to the HA+ fractions.

190 Figure 5: Percent of Total CTPs Captured as a function of Decreasing Label

50% 45%

40%

captured captured 35%

30%

CTPs CTPs 25% 20% 15%

10% Percent of total total of Percent Percent of total total of Percent 5% Number of total Number Captured CTPs 0% 1:1 1/3:1 1/9:1 1:1 1:1/3 1:1/9 1/9:1/9 Dilution of Primary and Secondary Label

Figure 5.17: Partitioning of CTPs after magnetic separation in the HA+ fraction. There was no significant effect of label reduction on the percent of CTPs captured in the HA+ fraction.

This data suggests that reduction of the primary label (HABP) can be achieved without sacrificing the number of cells and CTPs partitioning to the HA+ fraction. Reduction of the secondary label has a significant effect on the number of cells in the HA+ fraction, though not on the prevalence and number of CTPs. A trend toward an increased percent of CTPs captured in the HA+ fraction was seen in the sample with 1/3 HABP and normal concentrations of secondary labels, though this was not significant.

5.3.4 Discussion

Previously, hyaluronan (HA) has been evaluated as a novel positive marker for freshly aspirated CTPs, and has demonstrated enrichment of CTPs in the HA+++ cell fraction.

191 This study examines the effect of concentration of the magnetic label on CTP selection and magnetic mobility. We hypothesized that as the amount of reagent is reduced the percent of CTPs retained in the HA+ fraction will decrease.

In Aim 1, CTPs were enriched in the HA+++ cell fraction. One limitation of this study was the use of a single pass through the magnet instead of the three sequential passes through the magnet that produced the HA+++ fraction. A single pass through the magnet results in a less pure population, and a larger number of retained cells (albeit some nonspecifically trapped and retained). In order to examine the 6 different staining concentrations, marrow was split into smaller starting samples and processed. The resulting cell suspension after three passes through the magnet would not have resulted in enough cells to assay the positive fraction for CTPs and to test mobility using CTV. As a result, a single pass was used to assess these variables, noting that any chosen reduction in reagent concentration would have to be validated in the HA+++ population in future experiments.

The number of cells labeled in each tube was 150 million cells. Normally, 300 million cells are processed in 1.5 mL of volume. Since each of the 6 samples tested contained

150 million cells, the volume was scaled down to 0.75 mL to mimic the typical concentration of cells during labeling. An underlying assumption is that the concentrations used here will scale to larger volumes of marrow, with the larger total volume having no adverse effect on labeling. With consistent and thorough mixing and maintenance of the concentrations tested here, it is expected that the reagent amounts will scale appropriately.

192

A limitation of this study is its low power, as only three samples were tested. As a result, large decreases in CTP prevalence and partitioning are required to achieve significance.

Power analysis indicates that for a power of 0.8 and an N of 3, the detectable difference is a 2.6-fold change in prevalence (using a historical variability of 0.8 calculated from previous experimentation using a single pass through the magnet). In this case, the differences in CTP prevalence were expected to large as the labeling reagents were reduced. Instead, there was no significant difference in prevalence in the HA+ fractions, suggesting the amount of reagent normally utilized was well above the level needed to provide retention in the magnet with the retention time used (5 minutes). This was previously corroborated using a mathematical model to predict time required for retention of HA-labeled cells in the EasySepTM magnet using measurements made by CTV. The labeling was so strong that the average labeled cell was retained within 60 seconds in the magnet (Master‟s Project, paper in preparation).

CTV was used in this study to measure the magnetophoretic mobility of the magnetically labeled cells. However, the number of RBCs remaining in the cell suspension after only a single pass through the magnet made estimation of the mobility difficult using CTV. A

“gate” was applied to separate negative cells and RBCs from the positively labeled cells and the mean mobility of the positive cells in each sample was determined.

The total number of nucleated cells in the HA+ fraction decreased after reducing the amount of secondary antibody, but not after reducing the primary HABP. CTV analysis showed that the average magnetophoretic mobility of the HA-positive cells remained

193 fairly constant, indicating that either label could be reduced without adversely affecting the average positive mobility of the cells in the magnetic field.

Colony formation assays indicate that the prevalence of CTPs is not significantly different with reduction of both primary and secondary labels, even though the total number of cells captured in these fractions is reduced. This suggests that the CTPs are strongly magnetically labeled, as cells with weaker labeling would not be retained in the positive fraction as the amount of label is reduced.

There was no significant difference in the number of total CTPs captured in the HA+ fraction after reduction of either the primary or secondary antibody. This suggests the concentration of either or both the primary and secondary labels will maintain the total number of CTPs captured in the HA+ fractions. Reduction of the primary antibody by 1/3 resulted in the highest number of CTPs retained (trend not significant). Ongoing work in this project transfers this technology to a high capacity magnet. Reagent cost during scale up of the cell load can be expensive, and reduction of reagents will provide significant cost savings.

5.4 Effect of preincubation of labeling reagents to minimize the

number of labeling steps and incubation time required to

magnetically label the cells before separation in the magnet

5.4.1 Introduction

194 This experiment seeks to minimize the labeling steps and incubation time required to magnetically label the cells before separation in the magnet. Using the standard 3 pass protocol (3P), the HABP, EasySep TAC, and EasySep nanobeads are added sequentially, with an incubation period (60 minutes, 15 minutes, and 10 minutes, respectively) for each reagent. It would be preferable to use a one-step labeling and incubation, both to minimize the number of processing steps the cells undergo, and to reduce the overall time required for labeling.

Preincubation, if effective, would eliminate the wash step (5 min), TAC incubation time

(15 min) and bead incubation time (10 min), as well as minimizing the processing steps to which the cells are exposed (extra pipetting, mixing, centrifugation). Time for labeling would be reduced from approximately 1.5 hours to 1 hour.

We hypothesize that preincubation of reagents before labeling will not change the level of enrichment of CTPs.

5.4.2 Methods:

5.4.2.1 Cell Source

Bone marrow was aspirated from the humerus of 4 canine donors in 2 mL aliquots according to approved ACURO protocol #W81XWH-08-2-0034. Aspirates were collected in heparinized media. Two sequential density gradient separations (“buffy coats”) were used to deplete red blood cells. The resulting buffy coated marrow (BCM) cells were resuspended in phosphate buffered saline (PBS) with 2% fetal bovine serum

(FBS), 1 mM ethylenediaminetetraacetic acid (EDTA) and 2 units/mL heparin at 200

195 million cells/mL. The resulting cell suspension was split into two cell suspensions for comparison of the 3 pass (3P) and preincubation (INC) protocols.

5.4.2.2 Magnetic Separation the EasySepTM system

BCM cells were processed using the EasySepTM Magnetic Separation system (Stem Cell

Technologies #18543) on the basis of HA expression using a biotinylated G1-link protein

(hyaluronan binding protein, or HABP) (Calbiochem #385911), which binds strongly and specifically to HA.

3 Pass Labeling Protocol (3P)

Cells were labeled with 200 L/mL of FC blocker to prevent nonspecific uptake of

antibodies, followed by 6.7 L/mL of biotinylated HABP at 0.5 mg/mL for one

hour at room temperature. After removing excess HABP by centrifugation, the

EasySep anti-biotin tetrameric antibody complex was added at 200 L/mL and

allowed to incubate for 15 minutes at room temperature. The magnetic nanobeads

were subsequently added at 100 L/mL for a 10 minute incubation period.

Preincubation Labeling Protocol (INC)

Cells were labeled with 200 L/mL of FC blocker to prevent nonspecific uptake of

antibodies. For this protocol the HABP, TAC, and nanobeads, at the same

concentrations used in the 3P protocol, were preincubated together for 15 minutes,

then added to the cell suspension and incubated with the cells for 1 hour.

196 After increasing the total volume to 2.5 mL, the cells were placed in the EasySepTM magnet for 5 minutes, and the nonmagnetized population was decanted. Cells retained in the magnet for 3 sequential separations were labeled as the purified HA+++ population.

Cells that were not retained in the HA+++ fraction were sent bpack through the magnet.

Any cells retained on this subsequent pass were dubbed HA+, since the magnetic labeling of these cells was present, but not strong enough to enable consistent retention in the magnet. Cells that were unbound on all passes through the magnet were defined as HA-.

5.4.2.3 Cell Culture Conditions

HA+++, HA+, HA- from both 3P and INC protocols, as well as the BCM, were cultured under osteogenic media conditions, consisting of -MEM with 10% fetal bovine serum, 1 unit/mL penicillin, 0.1 mg/mL streptomycin, 10-8 M dexamethasone, and 50 g/mL ascorbate (added on the day of use). Each fraction was plated at a density of 250,000 cells

2 o per LabTek chamber (4.2 cm ), and cultured at 37 C at 5% CO2 with medium changes on

Days 2 and 3. Day 6 cultures were fixed with 1:1 acetone:methanol for 10 minutes and stained for nuclei (DAPI) and osteoblastic activity (alkaline phosphatase, AP)

5.4.2.4 Image Acquisition and Analysis

Culture wells were imaged and analyzed with the ColonyzeTM image analysis software to obtain CTP prevalence. LabTek culture chambers were scanned using a Spot RTSE 9.0

Monochrome-6 12 bit digital camera (Diagnostic Instruments Inc.) mounted on a Leica

DMRXA2 motorized microscope controlled by Metamorph (v6.3) imaging software. The

ColonyzeTM image analysis software is used to quantitatively identify colonies containing eight or more cells in a cluster using a standardized algorithm.

197

5.4.3 Results:

5.4.3.1 Cell Partitioning after MS processing

There was no significant difference in the partitioning of cells between the 3P and INC protocols (Figure 5.18). Cell loss was 20.7% in the 3P and 14.9% in the INC. There were significant differences in the percent of cells partitioning to each fraction. The percent of cells partitioning to the HA+++ fraction (3P: 1.2%, INC: 2.1%) was significantly different than the HA+ (3P: 4.7%, INC: 4.3%) and HA- (3P: 73.4%, INC: 78.6%) fraction. The

HA- fraction was also significantly different than the HA+ fraction.

p < 0.0001 100% N=4 p < 0.0001 90% Bars are standard deviation 3P 80% INC 70%

60%

50%

40%

30%

20% p = 0.0403

10% Percent of Total Cells in each Fraction each in Cells Total Percentof

0% Cells Lost HA+++ HA+ HA-

Figure 5.18: Cell partitioning after magnetic separation processing. There was no difference in cell partitioning between 3P and INC protocols.

5.4.3.2 CTP Prevalence

198 Figure 5.19 illustrates the prevalence of CTPs in each fraction using both the 3P and INC protocols. CTP prevalence, defined as the number of CTPs per million cells plated, is standardized to the buffy coated marrow aspirate (BCM). Both the 3P and INC HA+++ fractions are enriched in CTPs, 3.63-fold and 3.83-fold, respectively, over the BCM, while the HA- fractions are depleted in CTPs. There is no difference in CTP prevalence using the 3P versus the INC protocol.

GM 1 3 4 2 9

Prevalence (CTPs/million cells plated) cells (CTPs/million Prevalence 1 3.63 3.83 0 3P HA+++ 3P HA+ 3P HA- INC HA+++ INC HA+ INC HA-

Figure 5.19: CTP prevalence after magnetic separation standardized to the BCM. Both 3P and INC protocols resulted in significant increases in CTPs in the HA+++ fraction over the BCM. Both HA- fraction were depleted in CTPs.

5.4.3.3 Cells per colony

The number of cells per colony, an indicator of proliferation, can also be assessed using the ColonyzeTM software for each colony. Colonies from both 3P and INC protocols performed similarly, as seen in Figure 5.20. Colonies from the HA+++ fraction tended to

199 be larger (not statistically significant), while HA+ and HA- colonies were smaller than the

BCM (significant for the 3P HA+, INC HA+ and INC HA- fractions).

GM 1 3 4 2 1.6

1.4

1.2

0.8

0.6

Relative Cells per Colony Colony per RelativeCells 0.4 (StandardizedControl)BMA to 0.2

0 3P HA+++ 3P HA+ 3P HA- INC HA+++ INC HA+ INC HA-

Figure 5.20: Cells per colony, standardized to the BCM. There were no significant differences in the number of cells per colony between the 3P and INC protocols.

5.4.3.4 Alkaline phosphatase expression

Alkaline phosphatase area per cell number was calculated for each colony as a measure of osteogenic differentiation. There were no differences between the 3P and INC protocols or between the individual magnetic fractions.

5.4.3.5 Partitioning of CTPs after magnetic separation

Based on the initial prevalence of the BMA, the total number of CTPs present in the starting population can be predicted. After magnetic separation, the cell count for the each fraction multiplied by the CTP prevalence in that fraction to give the total number of

200 CTPs captured in each fraction. CTP loss was greater than the cell loss, with a CTP loss of 82.9% in the 3P and 75.2% in the INC.

There were significant differences in the percent of CTPs partitioning to each fraction within the 3P and INC fractions. The percent of CTPs partitioning to the INC HA- fraction (INC: 25.5%) was significantly higher than the 3P HA- fraction (3P: 9.7%), both

HA+ fractions (3P: 3.2%, INC: 4.0%), and both HA+++ fractions (3P: 4.2%, INC: 7.1%).

5.4.4 Conclusion

This experiment sought to minimize the labeling steps and incubation time required to magnetically label the cells before separation in the magnet. It would be preferable to use a one-step labeling and incubation, both to minimize the number of processing steps the cells undergo, and to reduce the overall time required for labeling.

There was no difference in cell or CTP partitioning between the 3P and INC protocols in the HA+++ fractions. Both the 3P and INC HA+++ fractions are enriched in CTPs, 3.63- fold and 3.83-fold, respectively, over the BCM, while the HA- fractions are depleted in

CTPs. Preincubation of label reagents is recommended for integration into future magnetic separation protocols.

5.4.5 References

1. Bøyum, A. Separation of leucocytes from blood and bone marrow

Scand. J. Clin. Lab. Invest. 21, suppl.97 1968

201 2. Dragoni, A., Angelini, A., Iacone, A., D‟Antonio, D., and Torlontano, G. Comparison of five methods for concentrating progenitor cells in human marrow transplantation. Blut.

60, 278, 1990.

3. Yeo, C., Saunders, N., Locca, D., Flett, A., Preston, M., Brookman, P., Davy, B.,

Mathur, A., and Agrawal, S. Ficoll-PaqueTM versus LymphoprepTM: a comparative study of two density gradient media for therapeutic bone marrow mononuclear cell preparations. Regen Med. 4, 689, 2009.

4. Kasten, P., Beyen, I., Egermann, M., Suda, A., Moghaddam, A., Zimmermann, G., and

Luginbuhl, R. Instant stem cell therapy: characterization and concentration of human mesenchymal stem cells in vitro. European Cells and Material 16, 47, 2008.

5. Villarruel S, Boehm C, Pennington M et al. The effect of oxygen tension on the in vitro assay of human osteoblastic connective tissue progenitor cells J Orthop Res 26(10):1390-

1397 2008

202 Chapter 6:

Defining a Method for Loading of Magnetically Separated Cells

onto Osteoconductive Scaffolds

6. Chapter Introduction:

Surgeons have many available methods to apply bone marrow-derived cells to a bone defect. Injection of bone marrow into the injured area of bone without any scaffold has been performed with some success.1-9 In recent years, cells are typically loaded onto selected osteoconductive scaffolds for repair of bone defects. The addition of a scaffold provides a stable and osteoconductive surface for CTPs, which are an adherent and tissue resident cell population in vivo, to attach and proliferate. For culture-expanded cell populations, usually mesenchymal stem cells (MSCs), the MSCs can be seeded onto 3D scaffolds and cultured in a bioreactor before implantation into the bone defect.10,11

Loading of non-culture-expanded cell populations, including bone marrow aspirates, can be achieved most simply by soaking the scaffold in the cell solution, dripping the cell suspension onto the scaffold, or applying a clot of cells.12,13 However, while these methods are simple and easy to perform intraoperatively, they typically do not allow for further preferential selection of CTPs into the scaffold.

Selective retention, developed in the Muschler lab, occurs when a cell suspension containing CTPs is passed through specific biomaterials and CTPs are preferentially retained on the scaffold surface more readily than other nucleated cells.13-15 This results

203 in a significant increase in the concentration and prevalence of CTPs within the matrix.

This does not require manipulation of the cells other than anticoagulation to prevent clotting. Moreover, when progenitor cells are retained more efficiently, they are not only enriched from contaminating RBCs, but also nucleated non-progenitor cells that compete with transplanted CTPs and compromise their survival and function in the wound site.

Selective retention is analogous to an affinity column and in clinical application the substrate or biomaterial matrix in the column also becomes the implantable delivery system for the retained cells.

This chapter evaluated methods, including selective retention methods, for loading magnetically separated HA-positive cells onto osteoconductive scaffolds. The preferred method of loading will be utilized in the canine femoral multidefect model in vivo in

Chapter 7.

6.1 Evaluation of CTP rentention efficiency and selection ratio on the

loading of HA+++ cells using selective retention (SR) on PLCL-TCP

scaffolds.

Objective: Determine if magnetic separation can be effectively combined with selective retention processing (SR) in a clinically feasible protocol (i.e. testing the hypothesis that

CTP-Os will retain their property of selective attachment to allograft surfaces after magnetic separation processing). Retention efficiency and selection ratio will be used as

204 outcome measures to determine the effectiveness of SR methods using the HA+++ cell population.

6.1.1 Methods:

6.1.1.1 Preparation of PLCL-TCP Scaffolds

Poly(L-lactide-co-e-caprolactone) (PLCL) scaffolds with addition of tri-calcium phosphate (TCP) were developed by Dr. Griffith (MIT Boston, MA) and fabricated by

Therics LCC (Morrisville, PA) using 3D printing. PLCL-TCP scaffolds were cylinders with waffle architecture (height = 15 mm; diameter = 9.9 mm) designed to facilitate the capture of cells during loading with marrow, and contain ~1 mm channels, ~70-80% wall porosity, and a wall thickness of ~0.5 mm. The maximum size of the pores was ~250 µm.

Each layer has a network of 1 mm macrochannels, reminiscent of a waffle layout, to facilitate cell migration, nutrient transfer and matrix production (figure 6.1).

Figure 6.1: PLCL-TCP “waffle” design scaffolds. 15 mm high x 9.9 mm diameter

These scaffolds were packaged and then sterilized at MIT in an Andersen Anprolene

AN74i Ethylene Oxide Sterilizer at room temperature for 12 h using standard protocol.

205 After purging the sterilizer with air for 2 h, the scaffolds were left to degas in a chemical fume hood for 24 h at room temperature before shipping to the Cleveland Clinic for implantation.

6.1.1.2 Marrow aspiration and preprocessing

Bone marrow was aspirated from 4 canines from the humerus in 2 mL aliquots. Aspirates were collected in heparinized media. Some heparinized bone marrow (BMA) was set aside for loading onto scaffolds and the rest was processed by density separation (“buffy coating”). Two sequential buffy coats were performed to remove the bulk of the red blood cells (RBCs), and 300 million buffy coated marrow cells (BCM) were resuspended to 1.5 mL with buffer consisting of phosphate buffered saline (PBS) with 2% fetal bovine serum (FBS) and 1 mM ethylenediaminetetraacetic acid (EDTA) with 2 units/mL heparin.

6.1.1.3 Labeling and Magnetic Separation using the EasySepTM system

BCM cells at 200 million cells/mL were labeled with 200 L/mL of FC blocker (#18553,

EasySepTM biotin labeling kit, Stem Cell Technologies, Vancouver, BC, Canada) to prevent nonspecific uptake of antibodies, followed by 20 L/mL of biotinylated hyaluronic acid binding protein (HABP, #385911 Calbiochem) at 0.5 mg/mL for one hour at room temperature. After removing excess HABP by centrifugation, the

EasySepTM anti-biotin tetrameric antibody complex (#18553, EasySepTM biotin labeling kit, Stem Cell Technologies, Vancouver, BC, Canada) was added at 200 L/mL and allowed to incubate for 15 minutes at room temperature. The magnetic nanobeads

(#18553, EasySepTM biotin labeling kit, Stem Cell Technologies, Vancouver, BC,

206 Canada) were subsequently added at 100 L/mL for a 10 minute incubation. After increasing the total volume to 2.5 mL, the cells were put in the EasySepTM magnet for 5 minutes, and the nonmagnetized population was decanted. Cells retained in the magnet for 3 sequential separations were labeled as the purified HA+++ population. Cells that were unbound on all passes through the magnet were defined as HA-.

6.1.1.4 Cell Culture Conditions

Cells in each fraction were counted using a Beckman Coulter ViCell XR Viability

Analyzer and Cell Counter (Beckman Coulter, #731050). HA+++ and HA- fractions, as well as unselected BCM, were cultured under osteogenic medium conditions, consisting of -MEM with 10% fetal bovine serum, 1 unit/mL penicillin, 0.1 mg/mL streptomycin,

10-8 M dexamethasone, and 50 g/mL ascorbate. Each fraction was plated at a density of

250,000 cells per LabTek (#177380, Nalge Nunc International, Rochester, NY) chamber

2 o (4.2 cm ), and cultured at 37 C at 5% CO2 with medium changes on Days 2 and 3. Day 6 cultures were fixed with 1:1 acetone:methanol for 10 minutes and stained for nuclei

(DAPI) and osteoblastic activity (alkaline phosphatase). Chamber were scanned and processed through the ColonyzeTM quantitative image analysis system (version 2.0).

6.1.1.5 Selective retention loading of cells onto PLCL-TCP scaffolds

HA+++ cells were loaded onto PLCL-TCP scaffolds using selective retention methods

(N=3). After increasing the total volume of the HA+++ cell suspension to 8.4 mL using complete medium, the cell suspension was split and loaded onto 2 PLCL-TCP scaffolds.

PLCL-TCP scaffolds (diameter: 10mm, height: 10 mm) were placed into a sterile cage

207 and loaded onto a syringe column for 4x loading. 4.2mL of HA+++ cell suspension was placed into the syringe on top of the cage. The cell suspension was pulled through the scaffolds at a rate of 25.6 mm/min. For comparison, heparinized bone marrow aspirate

(BMA) was loaded with the same volume (4.2 mL) of marrow at the same flow rate using selective retention (N=3). The starting samples and effluent from all scaffolds were counted for cells and assayed for CTPs.

6.1.2 Results:

6.1.2.1 Cell and CTP partitioning after EasySepTM magnetic separation

After magnetic separation using the EasySepTM system, cell counts was performed on each fraction. On average, 1.8 ± 0.6 % of the cells partitioned to the HA+++ fraction and

74.0 ± 11.4 % to the HA- fraction (p< 0.0001). The overall cell loss after magnetic separation was 24.2 ± 11.2%.

Cells from each fraction were cultured for 6 days and processed through the ColonyzeTM program. CTP-O prevalence was standardized to the BCM. The average enrichment in the HA+++ fraction was 2.3 ± 1.1-fold over the BCM, and 7.6 ± 4.9-fold over the HA- fraction (p=0.0256).

Calculation of the yield of CTPs (YCTP) during processing requires the total number of

CTPs in each fraction to be calculated and compared to the number of CTPs in the original starting sample (CTPO). The number of CTPs in a given fraction x (CTPX) can be

208 calculated based on the product of the total number of nucleated cells in sample x (Nx) and the CTP prevalence (PCTP) in sample x.

CTPX = NX * PCTP(X)

YCTP = [CTPHA+++ + CTPHA- ] / CTPO

Using these equations, the yield and partitioning of CTPs in each fraction was calculated.

While only 24.2± 11.2% of cells were lost, a mean of 59.9±26.4% of the CTPs were not found among the MS fractions. The HA+++ fraction only contained 1.8% of the nucleated

+++ cells, but 5.3 ± 2.8 % of the CTPs, consistent with the increase in PCTP in the HA fraction. The HA- contained 34.8 ± 24.2% of the CTPs, as well as the overwhelming majority of nucleated cells (74.0%). There was no significant difference in the number of

CTPs partitioning to the HA+++ fraction versus the HA- fraction.

6.1.2.2 Retention Efficiency after PLCL-TCP Scaffold Loading:

The HA+++ fraction was loaded onto PLCL-TCP scaffolds using selective retention (SR) methods (2 scaffolds per marrow sample). Cell counts were performed on the effluent samples and the retention efficiency was calculated.

The number of cells and CTPs retained in the material (NR and CTPR) is calculated from the effluent cell sample (NE and CTPE) subtracted from the original starting cell sample

(No and CTPo).

NR = NO - NE CTPR = CTPO - CTPE

The retention efficiency (RE) for CTPs or nucleated cells (CTPRE and NRE) is calculated as:

209 NRE = NSR/ NO CTPRE = CTPSR/ CTPO

+++ For scaffolds loaded with HA cells, NR(HA) = 34.4 ± 9.7%, and NR(BMA) = 58.7 ± 6.1% for the scaffolds loaded with BMA (p=0.0045). More nucleated cells were retained in the graft using the BMA cell suspension (figure 6.2).

Using the total number of CTPs predicted by the starting sample, the percent of total

CTPs retained in the grafts were calculated for each sample (CTPR). The mean CTPR(HA)

+++ was 45.4 ± 21.5% using HA cells, and CTPR(BMA) = 73.0 ± 5.7% using BMA

(p=0.0041). A higher percent of CTPs were retained in the PLCL-TCP scaffolds using

BMA, indicating that the HA+++ cell suspension was less efficient at retaining CTPs using this methodology (figure 6.2).

210 90 HA+++ p=0.0041 80 BMA p=0.0045 70

Retention Efficiency (%) Retention Efficiency 20

0 RE cell RE CTP

Figure 6.2: Retention efficiency RE (% of cells and CTPs retained in the graft) using BMA and HA+++ cells. The retention efficiency for both cells and CTPs was higher using BMA samples. Bars are standard deviation.

6.1.2.3 Selection Ratio after PLCL-TCP Scaffold Loading:

The selection ratio, defined as the ratio of CTP-Os to nucleated cells retained in the graft

The selection ratio (SR) for CTPs vs. nucleated cells is calculated as a ratio of retention efficiency:

SR = CTPRE / NRE

A selection ratio >1.0 implies positive selection or enrichment of CTPs with respect to other marrow cells. Both BMA and HA+++ cell populations performed similarly. The selection ratio for HA+++-loaded grafts was 1.6 ± 0.94 and SR for BMA-loaded grafts was

1.25 ± 0.13. The selection ratio was not significantly different between the HA+++ and

BMA samples. Since the percent of both cells and CTPs retained in the scaffold was

211 higher in the BMA samples, the ratio of CTPs to cells was proportionally similar to the

HA+++-loaded grafts.

While the selection ratio was similar between HA+++ and BMA-loaded scaffolds, the absolute number of cells and CTPs loaded and retained were very different. After magnetic separation, only 1.8% of the cells in the starting population partitioned to the

HA+++ population, with an average of 5.0 ± 1.6 million cells loaded onto each graft. A mean of 1.7 ± 0.67 million cells were retained while 756 ± 454 CTPs were retained. On average, this resulted in a final prevalence of 1 CTP per 1963 nucleated cells in the allograft.

In contrast, the same volume of BMA contained an average of 222 ± 68 million cells that were loaded onto MCA scaffolds. A mean of 131 ± 40 million cells were retained while

11929 ± 3098 CTPs were retained. On average, this resulted in a final prevalence of 1

CTP per 11153 nucleated cells in the allograft. The number of cells and CTPs retained were significantly different between BMA and HA+++ samples (all p < 0.0001), as was the number of nucleated cells per CTP (p = 0.0001).

6.1.3 Conclusion:

While HA+++ CTPs did attach to PCLC-TCP matrices after magnetic separation processing, the CTP retention efficiency of HA+++ cells was significantly lower than

BMA. Since the cell retention efficiency was also lower in the HA+++-loaded grafts, the selection ratio, which indicates the preferential retention of CTPs over other nucleated

212 cells, was not different than BMA-loaded grafts. The selection ratio for both BMA and

HA+++-loaded grafts (1.25 and 1.6, respectively) was lower than that previously seen with formulations of canine allograft, a typical scaffold used in bone grafting.15

The HA+++ cell suspension differs in several ways from BMA, which has been the prototype cell suspension used in the development of selective retention methods.

Alteration of the properties of the HA+++ cell suspension to simulate those of a BMA may provide better retention efficiency of CTPs (see Chapter Discussion).

6.2 Evaluation of CTP rentention efficiency and selection ratio

on the loading of MS-processed cells using manual selective

retention (MSR) or drip soaking (DS) on canine allograft

matrix.

Objectives: Determine an effective method for loading of MS-processed cells onto allograft matrix. Evaluate drip soaking and manual selective retention methods.

6.2.1 Methods:

6.2.1.1 Preparation of Mineralized Cancellous Allograft

Mineralized cancellous allograft (MCA) bone (canine) was used as the scaffold. MCA was prepared and packaged by the Musculoskeletal Transplant Foundation (Edison, NJ)

213 using canine bone from the proximal humerus, proximal tibia and contralateral proximal femur. Donor bone was sterily harvested from canine subjects used in prior experiments following euthanasia. Harvest, shipping and processing at MTF were performed using standardized methods that are in common use for preparation of commercially available mineralized cancellous bone matrix consistent with clinical guidelines established by the

American Association of Tissue Banks. Cuboidal chips were prepared with a dimension of roughly 3x3x3 mm to enable uniform packing. Sterile processing was maintained throughout with standard confirmatory cultures. No terminal sterilization was used.

6.2.1.2 Plasma and marrow collection and preprocessing

Bone marrow was aspirated from 4 canine donors (48 mL) according to approved

ACURO Protocol 08288003.09 W81XWH-08-2-0034, IACUC #08606. The mononuclear cell fraction was separated from bone marrow by centrifugation for 10 minutes (“buffy coating”). The procedure was repeated to better deplete the sample of red blood cells. Autogenous plasma was prepared for each subject by aspirating 20 mL of blood from the animal forelimb into a heparinized syringe and the plasma layer was removed after density separation (1600 RPM for 10 minutes).

6.2.1.3 Labeling and Magnetic Separation using the Hexapole Magnetic System

(HMS)

The buffy-coated marrow (BCM) was labeled on the basis of HA expression using a biotinylated G1-link protein (Calbiochem #385911) and separated using the HMS to

+ obtain HA W2 cells.

214

All BCM cells from each subject were resuspended in 4-5mL of separation buffer. Cells

TM were labeled with FC blocker (1 mL, #18553, EasySep labeling kit, Stem Cell

Technologies, Vancouver, BC, Canada) to prevent nonspecific uptake of antibodies.

Cells were incubated in biotinylated HABP (50 g in 0.5 mg/mL) for one hour at room temperature on a rotator. Excess HABP was removed by centrifugation at 350g for 10 minutes. An anti-biotin tetrameric antibody complex (1 ml, #18553, Stem Cell

Technologies) was added and incubated for 15 minutes at room temperature. Magnetic nanobeads (#18553, Stem Cell Technologies) were then added at 500 L for 10 minutes.

After increasing the total volume to 30-35 mL using Ca2+/Mg2+-free PBS with 2% FBS and 0.5 mM EDTA, the cells were transferred to a custom channel, placed into the HMS magnet for 20 minutes, allowing magnetized cells to be retained against the wall of the chamber and the HA- (nonmagnetized) cells to be decanted. Two additional washing steps were performed using 30-35 mL of buffer and separated in the magnet for an

- addition 10 minutes per wash. The HA W1 cell fraction was removed after the first wash,

- + and the HA W2 fraction was removed after the second wash. The HA W2 fraction that remained retained was then removed from the magnet, resuspended in 5-10 mL of buffer and pelleted for 10 minutes at 280g, resuspended in 2 mL of autogenous plasma, and divided into two 1 ml samples for loading onto allograft.

6.2.1.4 Loading of cells onto allograft scaffolds

+ The resulting suspension of HA W2 cells was loaded on 0.3g of chips of canine MCA using either a drip soaking (DS) method or manual selective retention (MSR).

215

Figure 6.3: Image of canine allograft chips used for loading of HA+++ cells. Chips are packed into sterile cages before loading with cell suspensions.

Two sets of MCA were placed in sterile cages (figure 6.3) and pre-wet with 1-2 mL of

+ canine plasma. The drip soaked scaffold was loaded with 1 mL of HA W2 cell suspension by slowly dripping the cell suspension over the top of the scaffold by pipette. The effluent was collected from bottom of the sterile dish. The MSR scaffold was loaded using manual selective retention, passing cells completely through the matrix. The caged

MCA was placed in a syringe tube and 1 mL of cell suspension was placed into the syringe tube over the top of the caged matrix. The cell suspension was slowly pulled through the matrix twice (pull through, push back up, pull through again) and the effluent was collected in the attached tube. Figure 6.4 illustrates the methods used.

216 Figure 6.4: Illustration of manual selective retention and drip soaking methods.

6.2.1.5 Cell Culture Conditions

The starting sample and both effluent samples were counted using a Beckman Coulter

ViCell XR Viability Analyzer and Cell Counter (Beckman Coulter, #731050). Each sample was cultured under osteogenic medium conditions, consisting of -MEM with

10% fetal bovine serum, 1 unit/mL penicillin, 0.1 mg/mL streptomycin, 10-8 M dexamethasone, and 50 g/mL ascorbate. Each fraction was plated at a density of

250,000 cells per LabTek (#177380, Nalge Nunc International, Rochester, NY) chamber

2 o (4.2 cm ), and cultured at 37 C at 5% CO2 with medium changes on Days 2 and 3. Day 6 cultures were fixed with 1:1 acetone:methanol for 10 minutes and stained for nuclei

(DAPI) and osteoblastic activity (alkaline phosphatase). Chamber were scanned and processed through the ColonyzeTM quantitative image analysis system (version 2.0). The

217 loaded allograft chips were cultured for 2 weeks and stained for alkaline phosphatase expression (N=2).

6.2.2. Results:

6.2.2.1 Cell and CTP Partitioning after HMS Magnetic Separation

After HMS separation, a cell count was performed on each fraction, presented in Figure

6.5. The majority of cells partitioned to the HA- fraction (42.1 ± 10.1%) after the first separation, similar to separation using the EasySepTM system. On average, 9.9 ± 5.2% of

+ the cells partitioned to the HA W2 fraction. The overall cell yield was 85.7 ± 6.3%.

Figure 5: Percent of cells in each fraction after HMS 50% 45% 40% 35% 30% p = 0.0124 25% 20% p = 0.0003 p = 0.0004 15% 10%

Percent of total population cell Percent 5% 0% + - - - HA+++HA W2 HA-HA HA--HA W1 HA---HA W2 N=4, standard error bars Figure 6.5: Cell partitioning after HMS separation. The HA- fraction contained a significantly higher percent of cells than all other fractions.

Cells from each fraction were cultured for 6 days and processed through the ColonyzeTM program. CTP prevalence is presented in Figure 6.6 standardized to the BCM. The

218 + average enrichment in the HA W2 fraction is 1.9-fold, and the HA-negative fractions are all significantly depleted in CTPs.

Figure 6: CTP Prevalence after HMS separation 3 3 *

2.52.5

2 2

1.51.5

1 1

0.50.5

Prevalence (CTPs/million cells plated) cells (CTPs/million Prevalence Prevalence (CTPs/million plated) cells Prevalence 0 0 + - - - HA+++HA+++HA W2 HA-HAHA- HA--HAHA--W1 HA---HAHA---W2 *CTP Prevalence standardized to BCM Figure 6.6: CTP Prevalence after HMS processing in each fraction, standardized to the BCM. The HA-negative fraction are all depleted in CTPs.

6.2.2.2 Retention Efficiency for Cell and CTPs after DS and MSR loading

+ The HA W2 fraction was loaded onto canine MCA chips using drip soaking (DS) or manual selective retention (MSR) methods. Cell counts were performed on the effluent samples and the retention efficiency was calculated. The NRE using DS (56.6 ± 27.7%) was not different than using MSR (47.1 ± 30.6%). Figure 6.7 shows the retention efficiency for cells and CTPs.

+ Effluent samples and the HA W2 starting sample were cultured and processed through the

ColonyzeTM program. The prevalence of CTPs in the effluent samples was not

219 significantly different from each other (DS: 150 CTPs/106 cells plated, MSR: 154

CTPs/106 cells plated). The effluent samples were significantly depleted in CTPs when

+ + 6 compared to the HA W2 fraction (HA W2 : 293 CTP-Os/10 cells plated), indicating that

CTPs were retained within the graft.

The retention efficiency (CTPRE) for CTPs was calculated for each sample. Using DS methods, CTPRE = 79.1 ± 13.3% and 78.4 ± 14.4% of CTPs were retained using MSR methods (not significantly different).

100% DS 90% MSR

80%

70%

60%

50%

40%

30%

20%

Retention efficiency for cells and CTPs (%) CTPs and cells for efficiency Retention 10%

0% RE cell RE CTP Figure 6.7: Retention efficiency for cells and CTPs after loading using drip soaking (DS) or manual selective retention (MSR) methods. There was no difference in RE using either method.

6.2.2.3 Selection Ratio after DS and MSR loading

The selection ratio is defined as the ratio of CTPs to nucleated cells retained in the graft.

Both drip soaking and manual selective retention loading methods performed similarly.

220 The selection ratio for DS-loaded grafts was 1.6 ± 0.7 and the ratio for MSR loaded grafts

+ was 2.3 ± 1.7. Once optimized for the HA W2 cell fraction, selective retention coupled with HMS separation may provide higher selection than drip soaking, however, manual

SR methods were not significantly different than drip soaking.

In preparation for in vivo studies, canine allograft cancellous bone matrix chips were loaded by either MSR or DS, cultured in vitro for 14 days, and subsequently stained for alkaline phosphatase (N=2) (figure 6.8). Unloaded allograft chips are shown for comparison. Comparable alkaline phosphatase activity was found using both methods.

Figure 6.8: Allograft chips loaded by either manual selection retention or drip soaking from 2 canine donors. Grafts were cultured for 2 weeks and subsequently stained for alkaline phosphatase. Unloaded allograft chips are shown for comparison.

6.2.3 Conclusions:

These data have demonstrated that CTPs that are enriched using MS processing retain their property of rapid and preferential attachment to allograft bone matrix. These

221 findings enable a clinical protocol beginning with MS processing to deplete non- progenitors, followed by further enrichment by selective retention.

Transplantation of CTPs at greater concentration, but with fewer competing and non- contributing non-progenitor cells, is expected to enhance the survival and performance of transplanted CTPs and result in a more rapid and effective bone healing response, particularly in large segmental bone defects.

6.3 Chapter Discussion:

This chapter described two studies that seek to develop a protocol for loading MS- processed HA- positive cells onto osteoconductive matrices in preparation for in vivo experimentation in the CFMD model.

In the first study, the objective was to test the hypothesis that CTPs will retain their property of selective attachment to allograft surfaces after magnetic separation processing and subsequently to determine if magnetic separation can be effectively combined with selective retention processing (SR). Retention efficiency and selection ratio were used as outcome measures to determine the effectiveness of SR methods using the HA+++ cell population.

While HA+++ CTPs did attach to PCLC-TCP matrices after magnetic separation processing, the retention efficiency of HA+++ CTPs was significantly lower than BMA

CTPs. Since the cell retention efficiency was also lower in the HA+++-loaded grafts, the

222 selection ratio, which indicates the preferential retention of CTPs over other nucleated cells, was not different than BMA-loaded grafts.

The selection ratio for both BMA and HA+++-loaded grafts was lower than that previously seen with some formulations of canine allograft.13,15 Muschler et al. loaded BMA onto

3x3x3 mm cubes of canine cancellous bone matrix (both demineralized and mineralized formulations) using selective retention methods. Using the mineralized cancellous allograft, the retention efficiency for cells was 49 ± 8%, and for CTPs was 57 ± 18%. The mean selection ratio was 1.2 ± 0.3.13 When BMA was loaded onto demineralized cancellous allograft, the retention efficiency for cells was 44 ± 19%, and for CTPs was 70

± 13%. The mean selection ratio was 1.9 ± 1.2.13 Another paper used SR methods to load

BMA onto demineralized canine allograft cortical bone matrix powder and resulted in a cell retention efficiency of 23 ± 8% and a CTP retention efficiency of 61± 14%. The selection ratio was 3.0 ± 1.5, which is the highest selection of CTPs over nucleated cells in any of the scaffolds tested.15

Using the same selective retention methodology, BMA loaded onto PLCL-TCP scaffold averaged 1.25± 0.13, while HA+++ cells loaded onto the same scaffold resulted in a selection ratio of 1.6 ± 0.94. While these results are in line with those using canine cancellous bone matrix (both demineralized and mineralized formulations), they are lower than using demineralized canine allograft cortical bone matrix powder.

223 The HA+++ cell suspension differs in several ways from BMA, which has been the prototype cell suspension used in the development of selective retention methods.

Alteration of the properties of the HA+++ cell suspension to simulate BMA may provide better retention efficiency of CTPs.

Previously, Muschler determined that a 2 mL volume of marrow aspirate contains ~20% contamination with peripheral blood.16 Using the cellular composition of bone marrow and blood from Williams Hematology17 and assuming this 20% contamination with peripheral blood, a bone marrow aspirate will contain approximately 43% granulocytes,

16% lymphocytes, 2% monocytes, 20% erythroid progenitors, and 14% myeloid progenitors, as well as small percentages of other cell types (CTPs, reticular cells, endothelial cells, etc…). Given that the average prevalence of CTPs in a BMA is 1 in every 20,000 nucleated cells16,18, for every 1,000,000 nucleated cells in a BMA there are

143,000,000 RBC and 50 CTPs. This illustrates the rarity of CTPs and also the importance of increasing the prevalence and concentration of CTPs before implantation in a bone defect.

The lower cell and CTP retention in the HA+++-loaded scaffolds may at least partially be a function of the differences in solution properties between a heparinized bone marrow aspirate and the HA+++ cell suspension. BMA contains 143 million RBCs for every million nucleated cells, and this population is largely absent in the HA+++ cell fraction.

Flow cytometric analysis indicates that less than 10% of the HA+++ cell fraction is glycophorin A positive (see Chapter 2 and 4). While the RBC population is not expected

224 to aid directly in CTP attachment, the large number of cells imparts a higher solution viscosity. The cells in the BMA are also suspended in autologous plasma, which contain proteins such as fibronectin (FN) and vitronectin (VN), that coat exposed scaffold surfaces, and may enable cell and CTP retention throughout the PLCL-TCP scaffold.

Both FN and VN have been found to enhance the attachment of MSCs and osteoblastic cells to biomaterial surfaces.19-21 The buffer used for the HA+++ fraction only contained

2% fetal bovine serum. In addition, the density of nucleated cells, RBCs, and plasma proteins in a BMA contribute to a higher overall viscosity than the buffer in which the

HA+++ cells are suspended. The mean density of cells in the BMA was 52.9 million cells/mL, while the density of HA+++ cell suspension was 1.2 million cells/mL. This lower viscosity affects the flow of the cell suspension through the matrix. This can be expressed by the Darcy Law, which describes convective flow through porous vertical beds.22 It relates the flow rate (Q) to the bed characteristics (length L, cross sectional area

A A, and the Darcy‟s Law coefficient KM), the fluid viscosity (, and the pressure drop (P

- P):

K A Q  M P A  P L

An increase in the solution viscosity will decrease the flow rate of the cell suspension through the scaffold pores, giving CTPs more time to adhere to the scaffold surface.

For the second experiment (loading canine allograft), the protocol was altered to better approximate the loading environment of a BMA for the HA-positive cells. The allograft was pre-wet with autologous canine plasma, and the HA-positive cells were resuspended

225 in plasma instead of buffer solution to provide plasma proteins to aid CTP retention in the allograft. The total volume for loading was decreased to 1 mL to increase the concentration of cells in the HA-positive fraction. The concentration of cells in the suspension increased from 1.2x106 cells/mL to 52.8x106 cells/mL. SR loading was performed manually (no set flow rate) due to the small volume. The cells were pulled through the scaffold 2x (pull through scaffold, push up and pull back through again) to allow extra opportunities for attachment to the allograft.

From the first experiment to the second, the scaffold was changed from PLCL-TCP to

MCA canine allograft chips. PLCL-TCP and MCA allograft were both tested in the

CFMD model without cells and the MTF MCA allograft significantly outperformed

PCLCL-TCP in the percent of new bone volume in the defect site. MCA allograft will be used as the preferred scaffold to assess the additive effects of bone marrow-derived cells, including MS-processed cells, in future experimentation in the CFMD model.

In the second study, the objective was to determine an effective method for loading of

+ HA W2 cells onto allograft matrix by evaluating two methods of loading: drip soaking and manual selective retention. In previous experiments using the EasySepTM magnet, the

HA+++ fraction was obtained after 3 sequential separations in the magnet. The protocol determined by the Zborowski lab with the HMS uses a single magnetic separation followed by 2 sequential wash steps, hence the change in nomenclature from HA+++ to

+ HA W2.

226 Overall, these findings demonstrate that it is possible to perform SR processing of marrow-derived cells and CTPs after MS processing (i.e. the cells and CTPs appear to retain their ability to attach preferentially to allograft matrix, resulting in an increase in concentration and prevalence of CTPs). However, drip soaking was chosen as the preferred method for loading MS processed cells in the first in vivo assessment in the

CFMD model. This choice reduced substantially the number of logistic variables (and potential sources of variation) in the methods used in this first in vivo assessment.

Loading using selective retention may provide a valuable step in further enriching the

CTP population in future experimentation, but the methodology needs to be further developed for retention of MS-processed CTPs, including defining the preferred flow rate, volume of cell suspension, and number of passes through the matrix. Future experimentation comparing these variables is necessary to provide the optimal CTP retention efficiency and selection ratio of MS-processed cells.

The biological basis for the attachment and retention of CTPs on some surfaces has multiple possible mechanisms. This includes non-specific interactions based on charge, hydrophobic or hydrophilic interactions; specific interactions between the material surface (including proteins or other bioactive molecules that may become bound or presented on the surface) and the cells (including both specific receptors that may be presented on the surface of the cell). In the setting of three dimensional or textured biomaterials, mechanical interaction between the cell and the surface may also play a role. The specific mechanism involved in cell-material interaction is likely multifactorial

227 and highly dependent on the state of the cell at the time that it is exposed to the surface and the biophysical properties of the specific material involved.

6.4 References:

1. Connolly, J., et al., Development of an osteogenic bone-marrow preparation. J Bone

Joint Surg Am, 1989. 71(5): p. 684-91.

2. Connolly, J.F., et al., Autologous marrow injection as a substitute for operative grafting of tibial nonunions. Clin Orthop, 1991. 266: p. 259-70.

3. Connolly, J.F., Injectable bone marrow preparations to stimulate osteogenic repair.

Clin Orthop, 1995. 313: p. 8-18.

4. Garg, N.K., S. Gaur, and S. Sharma, Percutaneous autogenous bone marrow grafting in 20 cases of ununited fracture. Acta Orthop Scand, 1993. 64(6): p. 671-2.

5. Healey, J.H., et al., Percutaneous bone marrow grafting of delayed union and nonunion in cancer patients. Clin Orthop, 1990. 256: p. 280-5.

6. Hernigou, P., et al., Percutaneous implantation of autologous bone marrow osteoprogenitor cells as treatment of bone avascular necrosis related to sickle cell disease. Open Orthop J, 2008. 2: p. 62-5.

7. Hernigou, P., et al., The use of percutaneous autologous bone marrow transplantation in nonunion and avascular necrosis of bone. J Bone Joint Surg Br, 2005. 87(7): p. 896-

902.

8. Hernigou, P., et al., Percutaneous autologous bone-marrow grafting for nonunions.

Influence of the number and concentration of progenitor cells. J Bone Joint Surg Am,

2005. 87(7): p. 1430-7.

9. Paley D, Young MC, Wiley AM, Fornasier VL, Jackson RW. Percutaneous bone

228 marrow grafting of fractures and bony defects. An experimental study in rabbits.

Clin Orthop Relat Res. 1986;208:300-12.

10. Meinel L, Fajardo R, Hofmann S, Langer R, Chen J, Snyder B, Vunjak-Novakovic G,

Kaplan D. Silk implants for the healing of critical size bone defects. Bone. 2005

37(5):688-98.

11. Depprich R, Handschel J, Wiesmann HP, Jäsche-Meyer J, Meyer U. Use of bioreactors in maxillofacial tissue engineering. Br J Oral Maxillofac Surg. 2008

46(5):349-54.

12. Takigami, H., et al., Bone formation following OP-1 implantation is improved by addition of autogenous bone marrow cells in a canine femur defect model. J Orthop Res,

2007. 25(10): p. 1333-42.

13. Muschler, G.F., et al., Spine Fusion Using Cell Matrix Composites Enriched in Bone

Marrow-Derived Cells. Clin Orthop Rel Res, 2003. 407: p. 102-118.

14. Muschler, G.F., Method for Preparing a Composite Bone Graft. US patent No.

5,824,084, October 20. 1998.

15. Muschler, G.F., et al., Selective Retention of Bone Marrow-Derived Cells to Enhance

Spinal Fusion. Clin Orthop, 2005(432): p. 242-251.

16 Muschler, G.F., C. Boehm, and K. Easley, Aspiration to obtain osteoblast progenitor cells from human bone marrow: the influence of aspiration volume. J Bone Joint Surg

Am, 1997. 79(11): p. 1699-709.

17 Williams Hematology 6th edition. Beutler E, Lichtman M, Coller B, Kipps T,

Seligsohn U. (eds.) McGraw-Hill Medical Publishing Division

229 18 Majors A.K., Boehm C.A., Nitto H., Midura R.J., Muschler G.F. Characterization of human bone marrow stromal cells with respect to osteoblastic differentiation. J Orthop

Res 1997. 15:546-57

19 Sawyer A, Hennessey K, Bellis S. The effect of adsorbed serum proteins, RGD and proteoglycan-binding peptides on the adhesion of mesenchymal stem cells to hydroxyapatite. Biomaterials 2007 28:383–392

20. Wilson CJ, Clegg RE, Leavesley DI, Pearcy MJ. Mediation of biomaterial-cell interactions by adsorbed proteins: a review. Tissue Eng 2005;11(1–2):1–18.

21. LeBaron RG, Athanasiou KA. Extracellular matrix cell adhesion peptides: functional applications in orthopedic materials. Tissue Eng 2000;6(2):85–103.

22. Ultman JS, et al. Biomedical Transport Phenomena 2007

230 Chapter 7:

In vivo transplantation of autogenous marrow-derived cells

following rapid intraoperative magnetic separation based on

hyaluronan to augment bone regeneration.

Authors: Caralla T 1,2, Joshi P1, Fluery S1, Luangphakdy V1, Shinohara K1, Pan W1,

Boehm C1, Bryan J4, Vasanji A4, Hefferan T3, Walker E1, Yaszemski M 3, Hascall V1,

Zboroswki 1, Muschler 1

1Cleveland Clinic, Cleveland, OH; 2Case Western Reserve University, Cleveland OH;

3Mayo Clinic, Rochester, MN; 4Image IQ, Cleveland, OH

Adapted from article submitted to Tissue Engineering Part A

7.1 Abstract:

Introduction: This project was designed to the hypothesis that rapid intraoperative processing of bone marrow based on hyaluronan could be used to improve the outcome of local bone regeneration if the concentration and prevalence of marrow-derived CTPs could be increased and non-progenitors could be depleted prior to implantation.

Methods: Hyaluronan was used as a marker for positive selection of marrow-derived

CTPs using magnetic separation (MS) to obtain a population of HA-positive cells with an increased CTP prevalence. Mineralized cancellous allogaft (MCA) was used as an

231 osteoconductive carrier scaffold for loading of HA-positive cells. The canine femoral multidefect model was used and 4 cylindrical defects measuring 10 mm in diameter and

15 mm in length were grafted with MCA combined with unprocessed marrow or with MS processed marrow that was enriched in HA+ CTPs and depleted in RBCs and non- progenitors. Outcome was assessed at 4 weeks using quantitative 3D microCT analysis of bone formation and histomorphological assessment.

Results: Histomorphological assessment showed a significant increase in new bone formation and in vascular sinus area in the MS-processed defects. Robust bone formation was found throughout the defect in both groups. Total mineral density in the defects, including new bone formation and residual allograft, as assessed by microCT, was greater in defects engrafted with MS processed cells but the difference was not statistically significant.

Conclusion: Rapid intraoperative MS processing to enrich CTPs based on hyaluronan as a surface marker can be used to increase the concentration and prevalence of CTPs.

MCA grafts supplemented with heparinized bone marrow or MS processed cells resulted in a robust and advanced stage of bone regeneration at 4 weeks. Greater new bone formation and vascular sinus area was found in defects grafted with MS processed cells.

These data suggest that MS processing may be used to enhance the performance of marrow-derived CTPs in clinical bone regeneration procedures. Further assessment in a more stringent bone defect model is proposed.

7.2 Introduction:

Bone regeneration in large bone defects and complex wounds remains an unsolved clinical challenge1. Osteoconductive scaffolds such as allograft cancellous bone can be

232 successful in small or contained defects. However, success rates drop as defect size increases and in settings compromised by prior scarring, bone loss, and vascular compromise. In more complex settings surgeons most often utilize autogenous cancellous bone or supplement an osteoconductive scaffold using aspirated bone marrow (BMA) or an osteoinductive agent such as BMP-2.2 BMA contains a heterogeneous population of osteogenic connective tissue progenitors (CTP-Os) which are thought to contribute to new bone formation.3 However, the prevalence of CTPs is low (1 CTP per 20,000 nucleated cells).4,5 Moreover, BMA also contains a large number of erythrocytes derived from contaminating peripheral blood, which do not contribute to a bone healing response.

Given the limitation of diffusion of oxygen and other nutrients into a bone grafting site larger than 1-2 mm in thickness, there is reason to expect that the survival and contribution of CTPs that are transplanted in this environment is compromised by competing non-osteogenic cells.3 As a result, methods to both increase the number of

CTP-Os in a wound site and decrease the number of non-progenitors is desirable and are hypothesized to increase the rate or extent of bone formation in a graft site.

Successful bone repair or regeneration in any clinical setting requires CTP-Os. While osteoconductive and osteoinductive materials may improve bone regeneration, only osteogenic cells generate new bone. CTPs are defined as tissue-resident stem or progenitor cells that proliferate to form a colony in vitro and can be induced to express one or more connective tissue phenotypes.3,6 CTP-Os represent the subset of CTPs that are capable of generating osteogenic progeny. Recent data suggests that all or almost all of the new bone formed at the site of a normal fracture is generated by local cells present

233 in the injured tissue site.7,8 As a result, in settings where the local CTP population is suboptimal, as in most complex defect sites, optimizing the bone healing response will require transplantation of CTPs. The most available sources of CTPs are autogenous cancellous bone or bone marrow harvested by aspiration. Many preclinical studies demonstrate improved graft performance when marrow-derived cells are added, even to small graft sites in young healthy animals. This strongly supports the premise that the

CTP-O population is suboptimal in most clinical settings and that optimal performance from any osteoconductive or osteoinductive material may require augmentation with

CTPs.9-14

Several markers have been proposed for use in identification and positive selection of osteogenic cells from bone marrow, including STRO-1,15 CD271,16 CD49a,17 CD146,18 and CD105.19 We have recently reported that the presence of hyaluronan on the surface of freshly isolated marrow-derived cells can also be used as a marker for positive selection.20

This project was designed to test the hypothesis that rapid intraoperative processing of bone marrow using HA as a marker for positive selection could be used to improve the outcome of local bone regeneration in vivo.

7.3 Methods:

7.3.1 Preparation of Allograft

Mineralized cancellous allograft (MCA) bone (canine) was prepared by the

Musculoskeletal Transplant Foundation (Edison, NJ) using canine bone from the

234 proximal humerus, proximal tibia and contralateral proximal femur. Harvest, shipping and processing at MTF were performed using standardized methods consistent with clinical guidelines established by the American Association of Tissue Banks. Cuboidal chips (3x3x3 mm) enabled uniform packing of 15-20 chips to fill each femoral defect site.

Sterile processing was maintained throughout with standard confirmatory cultures. No terminal sterilization was used.

7.3.2 Collection of heparinized Bone Marrow and Plasma

Bone marrow was aspirated in 2 mL aliquots from the proximal humerus of 10 canines.

Aspirates were mixed immediately in heparinized saline at a ratio of 1 mL of 1000 IU

Na-Heparin per 2 mL aspirate. A total volume of 51 mL of heparinized bone marrow

(hBMA) was harvested.

A 3 mL sample of hBMA was reserved while the rest of the sample (48 mL) was preprocessed. Two sequential “buffy coat” density separations were performed to remove erythrocytes (400g for 10 minutes on an Allegra 6 Beckman Coulter centrifuge).

These buffy coated marrow (BCM) cells were resuspended in 4-5 mL of separation buffer (Ca2+/Mg2+-free phosphate buffered saline with 2% fetal bovine serum and 0.5 mM ethylenediaminetetraacetic acid). Autogenous plasma was prepared by aspirating 20 mLs of blood from the forelimb into a heparinized syringe and density separated.

7.3.3 Magnetic Separation and Preparation of MCA Implants

235 BCM cells were processed using a custom built hexapole magnetic system (HMS) on the basis of surface presentation of hyaluronan (HA) using a biotinylated hyaluronic acid binding protein (HABP) (Calbiochem #385911).

9 TM Cells (range: 0.8 - 2x10 cells) were labeled with FC blocker (1 mL, #18553, EasySep labeling kit, Stem Cell Technologies (SCT)) then incubated in HABP (50 g in 0.5 mg/mL) for one hour on a rotator. Excess HABP was removed by centrifugation at 350g for 10 minutes. An anti-biotin tetrameric antibody complex (1 ml, #18553, SCT) incubated for 15 minutes followed by magnetic nanobeads (500 L, #18553, SCT) for 10 minutes. After increasing the total volume to 30-35 mL using separation buffer cells were separated in the HMS magnet for 20 minutes. Magnetized cells were retained against the channel wall and the HA- (nonmagnetized) cells were decanted. Two washing

- steps were performed (10 minutes/wash). The HA W1 cell fraction was removed after the

- first wash, and the HA W2 fraction was removed after the second wash. The retained

+ + HA W2 fraction was removed and pelleted for 10 minutes at 280g. HA W2 cells were resuspended in 2 mL of autogenous plasma and divided into two 1 mL samples for MCA loading.

Samples of 1 mL of heparinized bone marrow aspirate (hBMA) or 1 mL of magnetically

+ + separated HA W2 (MS/ HA W2) cells in plasma were loaded onto 0.3g of MCA using by drip soaking. Starting and effluent samples were counted for cells and assayed for CTPs.

The loaded grafts were implanted into the CFMD.

236 7.3.4 Canine Femoral Multidefect (CFMD) Model

The CFMD21-23 was conducted following approval from the Cleveland Clinic IACUC.

Ten skeletally mature purpose-breed female coonhounds (age 2 to 4, weight 32.1  1.8 kg) were utilized as animal subjects.

General anesthesia was initiated using pentothal (20 mg/kg IV) followed by intubation and closed isoflurane (0.5-3%) inhalation. The left hind limb was shaved, prepped and draped in a sterile fashion. A ten cm incision was made on the lateral aspect of the thigh, beginning four cm distal to the greater trochanter of the femur. Deep fascia was incised between the biceps femoris and the vastus lateralis and these muscles were bluntly separated. Femur was exposed in an extraperiosteal plane by elevating the vastus lateralis.

A customized drill template was fixed longitudinally on the lateral aspect of the femur with two 3.5 mm bicortical screws (Synthes USA, Paoli, PA). Four separate 10 mm diameter cylindrical defects were created in the lateral metaphyseal and diaphyseal femur.

Four identical 1.0 cm diameter and 1.5 cm long cylindrical defects were then created by sequential use of a circular starting trochar, a pointed drill and a flat finishing drill. A 1.0 cm diameter by 1.5 cm long stainless steel spacer was placed in each defect temporarily to allow hemostasis to be achieved without formation of a clot of wound blood in the space that was preserved for the scaffold. Appropriate scaffolds are then placed in each defect. The WBM and MS processed BM loaded scaffolds in each animal were assigned to either an “ABBA” configuration (5 subjects) or a “BAAB” configuration (5 subjects) to control for possible site or proximity effects.

237 The graft sites were protected mechanically using a stainless steel plate that was fixed to the femur using two screws placed into the screw holes created for initial fixation of the drill guide. The vastus lateralus was then replaced over the defect sites and under the biceps in an anatomic position. The wound was closed using 0 vicryl suture to reapproximate deep fascia, followed by 2-0 vicryl suture in subcutaneous and subcuticular layers, then staples in the skin.

Animals were allowed free access to food and water and return to full weight bearing with daily exercise. Euthanasia was performed 4 weeks after implantation using

Beuthanasia™ solution (5ml/ 5kg IV). The femur was explanted. Individual defect sites were separated using a band saw and fixed in 10% neutral buffered formalin. After 48 hours, the solution was replaced with 70% ethanol to prevent demineralization. Micro-

CT images were obtained of each graft site prior to processing for histologic assessment.

7.3.5 Assay for CTPs

Samples from each fraction, BCM, and hBMA were counted in a Beckman Coulter

ViCell XR Cell Counter (Beckman Coulter, #731050). Cells were plated at a density of

2 o 250,000 cells/chamber (4.2 cm ), and cultured at 37 C at 5% CO2 with medium changes on Days 2 and 3. Osteogenic medium consisted of -MEM with 10% fetal bovine serum,

1 unit/mL penicillin, 0.1 mg/mL streptomycin, 10-8 M dexamethasone, and 50 g/mL ascorbate. Day 6 cultures were fixed with 1:1 acetone:methanol for 10 minutes and stained for nuclei (DAPI) and osteoblastic activity (alkaline phosphatase, AP).24

Chambers were scanned and analysed using ColonyzeTM software, identifying colonies containing eight or more cells in a cluster.24,25 CTP prevalence, defined as the number of

238 CTPs/106 cells plated, was obtained. The number of cells/colony and AP area/cell number was quantified for each colony.

7.3.6 MicroCT Acquisition and Analysis

Quantitative assessment of each graft site was performed using micro computed tomography (MicroCT) and 3D segmental image post-processing using an eXplore Locus

MicroCT scanner (GE Healthcare, Milwaukee, WI) at 45 micron resolution.

Bone formation within the defect site was identified based on electron density at or above the density of native trabecular bone (1000+/- 100 HU) by a blinded operator. A “defect template”, 10 mm in diameter and 15 mm in length size was manually positioned.

Voxels within the defect above the threshold mineral density were segmented as bone.

Percent bone volume (%BV) was measured by counting the fraction of bone pixels within the defect. The pattern and density of bone formation in each defect was plotted for visualization by projecting %BV data as a 2D contour plot using a grey scale range of

0-30%BV where the X-axis indicates radial position from the center of the defect to the edge (range 0-5 mm) and the Y-axis represents vertical position within the defect from the bottom to the opening at the lateral cortex (range 0-15) (figure 7.1). Bone volume data were subdivided into regions of interest based on differences on local tissue environment. The pericortical (PC) region (between 8 to 12 mm from the bottom of the grafted site) represents the region of the graft site that is adjacent to the cortex, where the contribution of cells from perioteum, cortex and endosteum may contribute to new bone formation. In contrast, the intramedullary (IM) region (between 3 to 7 mm from the

239 bottom of the grafted site) is separated the endosteal sources of osteogenic cells and bounded only by cells from the marrow cavity, a different milieu. The PC and IM regions are further divided into a “center”, “middle” and “outer” region based on the radial position within the defect site. The scaffold performance in the center most region of the defect was projected to be the most discriminating with respect to likely efficacy in larger defects.

Figure 7.1: Micro-CT processing Technique. A. The 3D defect volume is defined using a standard 10 mm diameter x 15 mm long cylinder. B. Following segmentation, bone volume (BV) data is mapped onto a color 2D contour plot using a scale from 0-30% BV. C. The defect site is divided into regions for analysis. The Pericortical (PC) region and the Intramedullary (IM) region are defined based on vertical position from the bottom of the defect. Three regions of depth are defined based on radial distance from the center in millimeters: Center (C) = 0.25-1.75 mm, Middle (M) = 1.75 – 3.25 mm, and Outer (O) = 3.25 – 4.75 mm.

240

To minimize the confounding effects of variation in %BV that would be unrelated to scaffold performance, the region of interest for analysis of data within the defect sites was carefully defined. Variation due to non-radially oriented boundary effects at the top soft tissue interface and the bottom endocortical interface was minimized by eliminating the top and bottom 3-mm of each defect from analysis. The very center (0-0.25 mm) was removed from analysis due to the vanishing small sampling volume that is assessed as one approaches a radius of zero as this creates high variability in data near the center.

Finally, data from the very edge of each defect (4.75.-5.0 mm) was also removed due to minimize the potential for random inclusion of existing cortex in the analysis of the PC region due to any small error in positioning of the “defect template”.

76.3.7 Histological Analysis

Histology was performed at the Bone Histomorphometry Laboratory at Mayo Clinic.

Each specimen was dehydrated in a graded series of alcohols, and embedded in polymethylmethacrylate without decalcification. Using a Leica RM 2265 microtome, specimens were sectioned along the vertical axis in the middle of the defect site into 5 micron sections and stained with modified Goldner‟s Trichrome and a hematoxylin and eosin (H & E) stain. Sections were scanned using a NanoZoomer Digital Pathology

System (Hamamatsu) and analyzed using the software package IHC score (Bacus

Laboratories Inc., Lombard, IL).

The pericortical and intrameduallary area were analyzed across the entire defect width at

10 mm and 5 mm, respectively, from the bottom of the defect. The center, middle, and

241 outer regions were compared used the same radial dimensions as in microCT analysis.

Quantitative evaluation was performed with specific assessment of the area of new bone formation, woven bone, vascular sinus spaces, fibrous marrow, hematopoietic bone marrow, and residual allograft material.

7.3.8 Statistical Analysis

Due to the variation between subjects in CTP prevalence in the BCM sample, the prevalence for each sample was normalized to the BCM. These standardized values were log base 2 transformed in order to obtain a Gaussian distribution, and the means and 95% confidence intervals were calculated. Back transformation was used to provide the geometric mean and the 95% confidence interval, providing the relative magnitude of change in from the BCM. An ANOVA with post-hoc Tukeys test was used for comparison of fractions between different process steps.

For microCT data, a mixed model ANOVA was used to test for significant of differences

+ between scaffolds (hBMA or MS/ HA W2), sites (proximal or distal), depth regions (PC or

IM), and radial distance (center, middle, or outer). The effect of dog was accounted for as a random factor. All interactions were included. The analysis was performed with JMP

9.0 (SAS®, Cary, NC).

7.4 Results:

7.4.1 Comparison of hBMA and BCM

The mean concentration of cells in the hBMA sample increased from 61.5x106 cells/ml to

123.6x106 cells/ml, with a mean cell yield of 42.3±8.6%. The mean prevalence of CTPs

242 in the hBMA and BCM samples were not different (93.6±61.1 CTPs/106 cells vs. 89.5±

55.3 CTPs/106 cells, respectively), while the mean concentration of CTPs in the hBMA increased from 5554±3637 CTPs/ml to 11785±9225 CTPs/ml, with a CTP yield of

53.4±40.7%.

7.4.2 Comparison of the Magnetic Separation (MS) Fractions

- - The majority of cells partitioned to the HA fraction (67.5 ± 10.2 %). The HA W1 sample contained fewer cells than the HA- fraction (14.7 ± 4.3%, p<0.0001), and more cells than

- the HA W2 population (5.2 ± 2.5%, p= 0.0045). Only 7.8 ± 2.7% of the starting cell

+ - population was retained as HA W2 cells (p<0.0001 compared to HA )

CTP prevalence (PCTP) varied significantly between the different MS fractions as illustrated in Figure 7.2. Each fraction is standardized to the PCTP of the BCM (line at 1).

+ HA W2 PCTP was 2.0-fold greater than that in the BCM sample (95% CI: 1.2, 3.1), indicating enrichment of CTPs.

243

1 2 3 4 5 6 7 8 9 10 GM 7GM 7 7

6 6 6

5 5 5

4 4 4

3 3 3

2 2 2

1 1

Prevalence (CTPs/million Prevalence cells plated) Prevalence (CTPs/million Prevalence cells plated) Prevalence (CTPs/million Prevalence cells plated) 0 0 0 + - HA+++- HA- - HA-- HA--- HA+++HA+++HA W2 HA-HAHA- HA--HAHA--W1 HA--- HAHA---W2 Figure 7.2: CTP Prevalence after MS processing, standardized to the BCM. - - HA PCTP was 9.8 fold lower than the BCM (95% CI 0.0021, 0.21). HA W1 PCTP was 3.7 fold lower than the BCM (95% CI 0.082, 0.49).

- + A large difference was found when comparing PCTP in the HA population and the HA W2

+ - population. The ratio of PCTP in the HA W2 /HA population was 18.3-fold (p= 0.0001), indicating that the first separation was most effective in removing non-CTPs without displacing CTPs. The first separation resulted in the removal of 67.5% of all nucleated

- cells and only 8.5% of the CTPs. The first washing step (HA W1) resulted in the removal of 14.7% of the nucleated cells and only 4.7% of the CTPs, and the second washing step

- (HA W2) resulted in the removal of 5.2% of the nucleated cells and 6.6% of all of the

CTPs.

7.4.3 Comparison of Colony Metrics

244 The number of cells per colony was similar in all fractions. The area of AP staining per

- cell was similar in the colonies formed from all fractions, with the exception of the HA W1 samples where AP area per cell normalized to the BCM sample was 0.54-fold (CI 0.26,

0.87).

7.4.4 Accounting for the CTPs after MS processing

The cell yield after MS processing was 95.1 ± 11.2%. The total number of CTPs in each fraction can be calculated based on the product of the total number of nucleated cells in a given sample x (Nx) and the PCTP-O in sample x. Using these calculations, the yield and partitioning of CTPs was calculated. While a mean of only 4.9% of cells were lost, a

+ mean of 58.3% of CTPs were not found among the four MS fractions. The HA W2 fraction only contained 7.8% of the nucleated cells, but 21.9 ± 17.3% of the CTPs,

- - - consistent with the increase in PCTP. However, the HA , HA W1, and HA W2 contained only

8.5 ± 14.3 %, 4.7 ± 5.7% and 6.6 ± 12.1% of all CTPs, respectively. Since the BCM sample contained 53.4% of all of the CTPs that were present in the initial hBMA sample, this suggests a loss of 82.9 ± 13.5% of all CTPs during the cell preprocessing and MS processing.

7.4.5 Allograft Loading

The retention efficiency for cells (RECells) loaded onto the MCA matrix was calculated as

10,11 + cells retained/cells loaded. RECells for the MS/HA W2 population was 41.8 ± 18.4% and RECells for the hBMA population was 32.8 ± 11.6% (not different). However, the

245 + retention efficiency for CTPs (RECTPs) was significantly higher in the MS/HA W2 samples

(75.7 ± 22.2% vs. 49.0 ± 20.4%) (p=0.012).

The selection ratio (SR) (% of CTPs/% cells retained) in the MCA graft was 2.1 ± 0.9 for

+ MS/HA W2 samples, and 1.6 ± 0.7 in the hBMA samples, which was not significantly different. After loading hBMA, the prevelance of CTPs was 1 CTP for every 14934

+ nucleated cells in the graft. After loading MS/HA W2, there was 1 CTP for every 3565 nucleated cells in the graft (p=0.0283). The mean number of cells retained in the graft

+ was 16.1 ± 10.4 million cells for MS/HA W2 and 20.5 ± 8.7 million cells for hBMA. The

+ mean number of CTPs retained in the graft was 5813 ± 4113 for MS/HA W2 and 3284 ±

+ 3147 for hBMA. In MS/HA W2 grafts, the number of CTPs retained was higher and the number of nucleated cells was lower.

7.4.6 Outcome of In Vivo Implantation in the CFMD Model

There were no surgical or postoperative complications.

2-dimensional contour plots of mean % BV across all subjects are presented in Figure 7.3 illustrating the pattern, distribution and density of bone formation for all scaffolds. The highest percent bone volume was found at the periphery of the defect, with variable levels of penetration into the middle and central regions. In addition, in all cases, %BV was highest in the pericortical (PC) region of the defect, and lower in the intramedulary

(IM) region.

246

Figure 7.3: Results of MicroCT. A. 2D contour plots of mean %BV separated into PC and IM regions showing 0-30% BV. B. Historical data showing the empty defect in the CFMD at 4 weeks in the IM region (2-9 mm from the bottom of the defect). C. 2D plot illustrating %BV as a function of radial position. PC and IM regions are plotted for + hBMA and MS/HA W2 defect sites.

In Figure 7.3, the percent bone volume (%BV) for each group is presented as a 2D plot that illustrate the change in %BV relative to depth within the defect. Data for the PC

+ region alone and the IM region alone are presented separately. Both MS/HA W2 and hBMA loaded grafts showed robust bone formation at 4 weeks.

+ While the defects grafted with MS/HA W2 cells had a slightly higher %BV than sites grafted with hBMA, the difference was not statistically significant. Excluding the

+ residual allograft, the mean %BV for the overall defect for MS/HA W2 group was 33.1%

(95% CI: 29.1, 37.1), and mean %BV for hBMA group 30.7% (95% CI: 26.7, 34.7). The

247 mean %BV in the PC region (36.2% SE: 2.1%) was significantly greater than the IM region (27.7% SE: 2.1%) (p<0.0001).

7.4.7 Histological Analysis

Histological analysis showed an advanced stage of bone formation, remodelling, and marrow reconstitution in both groups. Active bone remodeling with formation and resorption occurred on both residual allograft and new bone surfaces. New bone was synthesized on the surface of the allograft throughout the defect. Marrow space remodeling with vascular spaces and patches of hematopoietic marrow were seen in both groups. Figures 7.4 and 7.5 illustrate representative images.

+ MS/HA W2-loaded allograft

H&E Goldners 0.63 x 1.25 x

10 x 10 x

248 + Figure 7.4: Histological staining of site grafted with allograft loaded with MS/HA W2 cells

Bone marrow aspirate (hBMA) - loaded allograft H&E Stain Goldners Stain 0.63 x 1.25 x

10 x 10 x

Figure 7.5: Histological staining of site grafted with allograft loaded with hBMA

Figure 7.6 provides histomorphometric data for soft tissue components. There was

+ significantly more sinus area in the MS/HA W2 group compared to the hBMA group

+ (p=0.024), indicating an increase in vascularity in the MS/HA W2 defect sites. There was

+ a trend toward reduced fibrous tissue and in the MS/HA W2 group, but this was not statistically significant. The IM region exhibited a higher fraction of hematopoietic marrow and vascular sinusoids and less fibrous tissue than the PC region in both groups.

249

Marrow Space

80% IM hBMA p < 0.0001 70% IM MS/HA+W2 PC hBMA 60% PC MS/HA+W2 WD hBMA 50% WD MS/HA+W2

40% p < 0.0001 30% p = 0.024 p < 0.0001

Percent of total area (%) area total Percentof 20%

10%

0% Sinus Fibrous marrow Hematopoietic marrow

Figure 7.6: Histomorphological analysis of soft tissue composition in defects grafted + with MS/HA W2 or hBMA loaded MCA. Bars are standard error.

Figure 7.7 illustrates the histological findings of increased bone volume and woven bone in the PC region, consistent with microCT data. The percent of new bone formed was

+ significantly higher in the MS/HA W2 defect sites (p=0.039). Residual allograft was

+ present in both groups with a trend towards less residual allograft in the MS/HA W2 group, but this was not statistically significant.

250

Bone Distribution p < 0.0001 40% IM hBMA p = 0.039 35% IM MS/HA+W2 PC hBMA 30% PC MS/HA+W2 WD hBMA 25% WD MS/HA+W2

20%

15%

Percent of total Area (%) Area total Percentof 10% p=0.021 5%

0% New bone Residual allograft Woven bone

Figure 7.7: Histomorphometric analysis of bone composition in defects grafted with + MA/HA W2 or hBMA loaded MCA. Bars are standard error.

7.5 Discussion

These data support the hypothesis that rapid intraoperative processing of bone marrow using HA as a marker for positive selection can improve the outcome of local bone regeneration in vivo. Histomorphometric assessment demonstrated that the volume of new bone formation was significantly greater when MS processing was used. The area of vascular sinusoids was also significantly greater compared to unprocessed marrow aspirates. These findings suggest that MS processing may be a useful clinical method for enhancing the performance of marrow-derived cells.

251 The MS processing strategy used here is designed to address the profound challenge of cell survival following transplantation into defects greater than 1-2 mm thick, where imbalance between diffusion of oxygen and the metabolic demand of cells allow few cells to survive in deeper regions of the graft.3 A logical strategy for improving bone regeneration is to increase the concentration of bone forming cells. However, in larger defects, survival of these cells may be limited unless methods are also designed to limit

(deplete) the number of cells that do not contribute to new bone formation and/or cells which may impair or inhibit the survival and differentiation of CTPs, either by the release of inflammatory or pro-apoptotic cytokines or by cell death leading to secondary inflammation (and associated metabolic demand).

+ MS processing used here to isolate cells in the MS/HA W2 was successful in both increasing the concentration of CTPs among transplanted cells and in depleting non- progenitors and red blood cells. This finding was consistent with the finding of in a prior study.20

The outcome of MS processing of canine marrow in the current study was consistent with our prior experience with human marrow, with one exception. When initially reported in human marrow, we found that the HA+++ population that was isolated had a 3.4 fold higher prevalence of hCTPs, but also that the colonies formed by HA+++ hCTPs were more proliferative and exhibited a higher amount of AP staining than hCTPs in BCM.20

+ In the current study using canine marrow, the cCTPs that were present in the MS/HA W2 demonstrated a 2.0 fold greater prevalence, but there was no difference in proliferation

(cells per colony) nor in AP staining between colonies formed by cCTPs in the

252 + MS/HA W2 fraction when compared to colonies formed by cCTPs remaining in other fractions. It is not clear if the difference in this finding represents a difference in biology of marrow-derived CTPs between humans and canines, or if these findings represent differences in processing protocols. The processing methods used in the prior work in humans involved three separate steps of magnetic separation, resuspension and magnetic separation to provide an HA-, HA+, and HA+++ population using a much smaller volume of marrow and a smaller magnet. In contrast, MS processing in the current study involved only a single step of magnetic separation followed by two washing steps to

+ provide a HA W2 population. The labeling protocol was similar in both cases as were the field strength of the magnets used for separation (0.64 tesla for the prior EasySepTM system used in human marrow, and 0.74 tesla for the HMS magnet used here). Further work will be needed to distinguish between species specific and process specific differences.

The small scale EasySepTM magnet can process 2.5 mL of bone marrow in a single procedure, however, up to 48 mL of bone marrow is routinely aspirated and added to bone grafts in the settings of large bone defects, which are more likely to require augmentation with CTPs. In order to test the performance of HA-positive cells in the biologically relevant canine femoral multidefect model, processing of larger volumes of marrow was required. Scale up of magnetic separation was achieved using the hexapole magnetic system (HMS), developed in the Zborowski laboratory, which provides a portable permanent magnet with the capacity to separate 50 mL of marrow in a single procedure. The capacity of the HMS is 20 times that of the EasySepTM magnet.

253

Despite the positive in vivo effect of MS processing, an analysis and accounting of the fate of cells and CTPs during the processing suggests that there is significant room for process improvement. The first step in the separation, capture of HA+ cells and decanting of HA- cells that were not retained, was the most effective step. This resulted in the depletion of 67.5% of all of the nucleated cells with a CTP loss of only 8.5%. The first washing step resulted in the removal of 14.7 % of all of the nucleated cells and only

- 4.7 % of the CTPs in the HA W1 fraction. This fraction had a significantly lower CTP prevalence than the BCM, suggesting that this step is also effective at removing more nucleated non-progenitor cells than CTPs. However, in the second wash step the CTP

- prevalence in the HA W2 fraction was not significantly different than BCM, indicating that the concentration of CTPs removed in the second wash was higher than that removed in the first separation or the first wash step. Elimination of this second wash step would have preserved an additional 6.6% of CTPs.

In addition to opportunities to improve efficiency in the MS processing steps, there is also room for improvement in steps prior to magnetic separation during. A mean of 45% of all CTPs in the initial sample were lost during the process of preparing the BCM sample. These CTPs may have been lost due to adherence of CTPs to the surface of the containers use during preparation. A reduction in contact with potentially adherent surfaces may improve future yield. In addition, many CTPs have a cell density that is not low enough to allow separation into a buffy coat layer, and are retained in the red cell fraction during density separation. Alternative methods for preparation of the starting sample to remove RBCs while preserving CTPs need to be explored.

254

It is considered most likely that CTPs were lost due to their intrinsic adherence to surfaces during processing. However, it is also possible that the labeling procedures and processing in suspension resulted in a change in cell state that diminished adherence or proliferative capacity, reducing colony forming efficiency (CFE). MS-based isolation of subsets of cells may have resulted in the partitioning of non-osteogenic “helper” cells that were no longer co-cultured with CTPs, causing a reduction in CFE. These cells may provide cytokines and/or cell signaling that aid in CTP colony formation. Baksh et al. cultured BM-derived progenitor cells with conditioned media from hematopoietic cell cultures and found a significant increase in CFU. They postulated that the hematopoietic cell population secreted important regulatory molecules that support the growth and survival of mesenchymal progenitor cells.27 Previously, we noted that hematopoietic progenitor cells do not co-localize with CTPs in the HA+++ fraction.20 Perhaps the partitioning of these cells away from the HA-positive CTPs negatively affected their true potential for growth and colony formation.

In addition, an inherent limitation of the ColonyzeTM software is the underdetection of

+ colonies in slides that have a high prevalence of CTPs, like those in the HA W2 fraction.

Closely spaced colonies in samples with high prevalence may be indistinguishable from a single larger colony. The enrichment levels quantified and presented here are

+ conservative and likely underestimate the true prevalence of CTPs in the HA W2 fractions.

255 Several steps are needed to refine the protocol and conditions to prevent the loss of CTPs.

Steps can be taken to minimize the processing steps the cells are exposed to, including preincubation of labeling reagents and the elimination of the second wash step.

Preincubation will eliminate centrifugation and mixing steps as well as reduce the time required for labeling. Elimination of the second wash step will eliminate a processing

+ step and retain additional CTPs. Finally, after the separation is complete and the HA W2 cells removed, a trypsin digestion on the channel and the labeling tube should be performed to check for adherent CTPs.

The fact that CTPs retain their ability to adhere to MCA following MS processing is an important finding. This finding allows cell populations that are processed using MS methods to be combined with MCA and other potential scaffold materials and transplanted into a graft site on the surface of a biomaterial. Adherence to the surface of a biomaterial greatly enables the mechanical manipulation and placement of the processed cells into a defect and increases the likelihood that the transplanted cells will be retained at the site. The high CTP retention efficiency during loading indicates that

MS processing resulted in no compromise of CTP adherence properties. Moreover, the selection ratio greater than 1.0 also indicates that CTPs were preferentially retained on the allograft chips using simple drip soaking loading methods (i.e. CTPs were more likely to be retained than non-CTPs), providing an additional step in which non-progenitors can be further depleted prior to implantation.

Histological analysis of the defect sites after 4 weeks showed an advanced stage of bone formation, remodelling, and marrow reconstitution in both groups. Most of the donor

256 allograft was resorbed by 4 weeks, as residual allograft accounted for only 3.5%

+ (MS/HA W2 grafts) and 4.7% (hBMA grafts) of the total defect area. New bone was synthesized de novo and on the surface of the allograft throughout the defect, with a higher percent of new bone in the pericortical area for both graft types. There was

+ significantly more new bone in MS/ HA W2 defect sites than in hBMA defects as assessed by quantitative histomorphometry.

Marrow space remodeling with vascular spaces and patches of hematopietic marrow were

+ seen in both groups. More vascular sinus area was detected in MS/ HA W2 defect sites than in hBMA sites, indicating an increase in vascularization in the defects implanted

+ with MS/ HA W2 cells. There was no evidence of an inflammatory response in defect sites

+ implanted with MS/ HA W2 cells with attached magnetic beads in the H&E and Goldner stained slides.

The CFMD model used here has proven to be an effective system for rigorous and relatively rapid objective comparison between bone grafting materials and strategies in the setting of a modestly large defect in a large animal. Directly comparing two grafting techniques in a single animal controls for variation between subjects and increases statistical power by offering paired comparisons, which in turn reduce the number of subjects needed to address a given comparison, sparing animal lives.21 The defects are large enough so that the center of the defect will become profoundly hypoxic within 2-3 days following surgery. As a result, it would be expected to be sensitive to changes in cell transplantation strategies that might limit hypoxia and improve the survival of

257 transplanted cells. Previous experience has demonstrated that the canine proximal humerus is a reliable source of hematopoietic bone marrow containing osteoblastic progenitors, with yields of nucleated cells and CTP-Os that are comparable to human iliac crest aspirates.26

The CFMD model proved to be a limitation in this study in one respect. Bone regeneration in the MCA grafts supplemented with heparinized bone marrow was so robust compared to prior scaffolds used that it left relatively little room for improvement.

This resulted in a potential “ceiling effect” in which the magnitude of effect from modification of the cellular environment within the defect site might be underestimated.

Even though a positive effect was found, this finding suggests that if the benefits of MS processing of marrow is to be confirmed and if the value of making further improvement in cell yield and selection is to be tested, this assessment should be preformed in an even more stringent model involving a larger defect and perhaps a soft tissue environment that is compromised with respect to scarring and tissue loss and therefore more representative of the most challenging clinical defects. Further assessment using other less robust scaffolds as substrates may be considered, recognizing this potential limitation.

7.6 Acknowledgements:

We would like to acknowledge MTF for proving the MCA allograft matrix, the histology team at the Mayo Clinic for staining and histomorphometric analysis, Rick Rozic for

ColonyzeTM imaging and microCT scanning, and Tess Henderson for microCT analysis.

258 7.7 References:

1. Bishop G.B., Einhorn T.A. Current and future clinical applications of bone

morphogenic proteins in orthopaedic trauma surgery. Internal Orthpaedics 2007.

31:721-727

2. Rockwood and Green 6th Ed. “Fractures in Adults” Editors R. Bucholz, J

Heckman, C. Court-Brown. Lippincott, Williams, & Wilkins 2006. Philadelpia

PA.

3. Muschler G, Nakamoto C, Griffith L. Engineering principles of clinical cell-

based tissue engineering. J Bone Joint Surg 2004. 86A(7):1541-1558

4. Muschler G., Boehm C., Easley K. Aspiration to obtain osteoblast progenitor

cells from human bone marrow: the influence of aspiration volume. J Bone Joint

Surg Am 1997. 79:1699-709

5. Majors A.K., Boehm C.A., Nitto H., Midura R.J., Muschler G.F. Characterization

of human bone marrow stromal cells with respect to osteoblastic differentiation. J

Orthop Res 1997. 15:546-57

6. Muschler G., Midura R. Connective tissue progenitors: practical concepts for

clinical applications. Clin Orthop 2002. 395:66-80

7. Shinohara K, Greenfield S, Pan H, Vasanji A, Kumagai K, Midura RJ,

Kiedrowski M, Penn MS, Muschler GF. Stromal cell-derived factor-1 and

monocyte chemotactic protein-3 improve recruitment of osteogenic cells into sites

of musculoskeletal repair. J Orthop Res. 2011 29(7):1064-9

8. Boban I, Barisic-Dujmovic T, Clark S. Parabiosis Model Does Not Show

Presence of Circulating Osteoprogenitor Cells. Genesis 2010 48:171-182

259 9. Hernigou, P., Poignard A, Beaujean F, Rouard H. Percutaneous autologous bone-

marrow grafting for nonunions. Influence of the number and concentration of

progenitor cells, J Bone Joint Surg Am, 2005. 87 (7):1430-7

10. Muschler, G. F., Nitto H., Matsukura Y, Boehm C., Valdevit A., Kambic H.E.,

Davros W.J., Powell K. & Easley K. Spine fusion using cell matrix composites

enriched in bone marrow-derived cells, Clin Orthop Rel Res, 2003. 407:102-18

11. Muschler, G. F., Matsukura Y., Nitto H., Boehm C.A., Valdevit A.D., Kambic

H.E., Davros W.J., Easley K.A. & Powell K., Selective retention of bone marrow-

derived cells to enhance spinal fusion, Clin Orthop, 2005. (432):242-51

12. Tiedeman, J. J., Connolly J.F., Strates B.S. and Lippiello L., Treatment of

nonunion by percutaneous injection of bone marrow and demineralized bone

matrix. An experimental study in dogs, Clin Orthop, 1991. 268:294-302

13. Connolly, J. F., Injectable bone marrow preparations to stimulate osteogenic

repair, Clin Orthop, 1995. 313:8-18.

14. Hernigou, P., Poignard A., Manicom O., Mathieu G., & Rouard H. The use of

percutaneous autologous bone marrow transplantation in nonunion and

avascular necrosis of bone, J Bone Joint Surg Br, 2005. 87 (7):896-902

15. Stro-1 Simmons P, Torok-Storb B Identification of stromal cell precursors in

human bone marrow by a novel monoclonal antibody. Blood. 1991 78:55-62

16. CD271 Quirici N, Soligo D, Bossolasco P et al. Isolation of bone marrow

mesenchymal stem cells by anti-nerve growth factor receptor antibodies. Exp

Hematol. 2002 30: 783-91

260 17. CD49a Deschaseaux F, Chambord P. Human marrow stromal cell precursors are

alpha 1 integrin subunit-positive. J Cell Physiol 2000 184: 319-25.

18. CD146 Crisan M, Yap S, Casteilla L et al. A perivascular origin for mesenchymal

stem cells in multiple human organs. Cell Stem Cell 2008. 3:301–313

19. CD105 Boriet N, Rapatel C, Boisgard S et al. CD34+ CDw90(Thy-1)+ subset

collocated with mesenchymal progenitors in human normal bone marrow

hematon units is enriched in colony forming unit megakaryocytes and long-term

culture-initiating cells. Exp Hematol 2003 31:1275-83

20. Caralla T, Boehm C, Hascall V, Muschler G. Hyaluronan as a novel marker for

rapid selection of connective tissue progenitors. Submitted to J Ortho Res

21. Takigami H, Kumagai K, Latson L, Daisuke Togawa D, Bauer T, Powell K,

Butler RS, Muschler GF Bone Formation following OP-1 Implantation Is

Improved by Addition of Autogenous Bone Marrow Cells in a Canine Femur

Defect Model. J of Orthop Research, 2007.25:1333-42

22. Muschler GF, Raut VP, Patterson TE, Wenke JC, Hollinger JO The Design and

Use of Animal Models for Translational Research in Bone Tissue Engineering

and Regenerative Medicine. Tissue Engineering, Part B Rev. 2010. 16(1):123-45.

23. Raut VP, Patterson TE, Wenke JC, Hollinger JO, Muschler GF Assessment of

Biomaterials: Standardized In vivo Testing. An introduction to biomaterials. S.A.

Guelcher, J.O. Hollinger (eds), CRC Press, 2010

24. Villarruel S, Boehm C, Pennington M et al. The effect of oxygen tension on the in

vitro assay of human osteoblastic connective tissue progenitor cells J Orthop Res

2008. 26(10):1390-1397

261 25. Powell K, Nakamoto C, Villarruel S et al. Quantitative image analysis of

connective tissue progenitors. Anal Quant Cytol Histol 2007. 29(2):112-121

26. Muschler GF, Huber B, Ullman T, Barth R, Easley K, Otis JO, Lane JM. Evaluation of

bone-grafting materials in a new canine segmental spinal fusion model J Orthop Res.

1993 Jul;11(4):514-24.

27. Baksh D, Davies J, and Zandstra P. Soluble factor cross-talk between human bone

marrow-derived hematopoietic and mesenchymal cells enhances in vitro CFU-F

and CFU-O growth and reveals heterogeneity in the mesenchymal progenitor cell

compartment Blood 2005. 106: 3012-3-19.

262 Chapter 8: Future Work

8.1 Summary

8.1.1 HA as a novel marker for CTPs

In Chapter 3, hyaluronan was used as a novel marker for selection of CTPs from a fresh marrow aspirate. Data demonstrates that at least one subset of the CTPs is enriched by selecting for cell-associated HA (3.4-fold compared to unselected bone marrow). These cells are significantly more proliferative, as assessed by the number of cells per colony, and more readily differentiate to an osteoblastic phenotype in vitro when compared to unselected marrow. In contrast, both myeloid and erythroid hematopoietic progenitor cells do not segregate with CTPs and remain in the HA- fraction. Only 2.7% of cells were retained in the HA+++ fraction, eliminating the majority of non-osteogenic competing cells. This is the first study evaluating the use of hyaluronan as a potential marker for marrow-derived CTPs. The highly proliferative HA+++ cell population may provide a useful cell source for bone regeneration.

8.1.2 HA as a potential niche component for CTPs in vivo

The enrichment of CTPs based on HA, an extracellular matrix molecule, suggests that

HA represents a component of the niche surrounding at least one osteogenic progenitor cell population in human bone marrow. In Chapter 3, data demonstrated that CTPs retain

HA from their in vivo niche rather than synthesize it after aspiration. There was no difference in mean CTP enrichment after magnetic separation between samples processed on ice (during which the HAS enzymes are inactivated) and at room temperature.

263 The anatomic origin of the HA+++ CTPs isolated in this study is unknown, however, histological analyses show that HA is observed in the ECM around two subsets of cells in human marrow. In sections of rat tibia, strong HA staining is present in the perivascular area. HA staining was also observed surrounding isolated stromal cells in the marrow space.1 Echoing these results, sections of human bone marrow show staining of HA in stromal extracellular matrix, perivascularly around the marrow sinusoids, and on the endosteal surface of bone.2-4 Evidence supports a perivascular niche for at least one subset of osteogenic cells in bone marrow (see Chapter 1). CD146, a pericyte marker, has been used in combination with other cell surface markers to isolate osteogenic progenitor cells. CD45-/CD146+ cells are enriched in CTPs, as are STRO-1BRT/CD146+ cells (see

Chapter 2).5,6

8.2 Future work:

Further research to uncover the in vivo location of HA+++ CTPs and cell surface marker profile is an active area of research. Many researchers have suggested subsets of CTPs reside in vivo in a perivascular niche and can be characterized by their expression of

CD146. Histological analysis of HA shows strong staining in the perivascular area, the known site of CD146+ pericytes in bone marrow. Are the CTPs isolated using HA-based separation derived from the perivascular area? Are they the same progenitors that other researchers have isolated using CD146? Future studies to examine human iliac crest bone cores will determine to what extent freshly isolated cells express CD146 and also retain

HA on their surface. Dual color histology can be performed for HA and CD146 to determine the relative prevalence of HA+ , CD146+, or HA+/CD146+ cells and to

264 determine if these variations in markers represent cell sets with the same or different biological potential.

Preliminary data shows that CD146 can be used as a marker for CTPs and the level of enrichment of CTPs from fresh marrow aspirates is similar to that observed for HA-based selection (N=4). Using EasySep magnetic separation, cells from the same patient were labeled separately for HA and CD146 and magnetically separated. Following selection for hyaluronan, 1.8 ± 1.1% of the starting population partitioned to the HA+++ population and 2.4 ± 2.1% of the starting population partitioned to the CD146+++ population. The level of enrichment was 1.8-fold for the HA+++ fraction and 2.9-fold for the CD146+++ fraction over the buffy coated marrow. HA- and CD146- fractions were depleted in CTPs.

After 6 days of culture, colonies were stained for DAPI, HA, and CD146 and analyzed using ColonyzeTM. Colonies co-expressed HA and CD146 in the HA+++ and CD146+++ fractions (figure 8.1). Additional patients need to be added to this study to determine the reproducibility and significance.

HA+++ fraction CD146+++ fraction

265 Figure 8.1: Co-expression of CD146 and HA on a colony after 6 days of culture. A colony derived from the HA+++ fraction is on the left and the CD146+++ fraction on the right. Cells are stained for their nuclei with DAPI (blue), HA (red), and CD146 (green).

While dual expression of CD146 and HA in CTP-founded colonies is intriguing, cells can up- or down-regulate cell surface markers during culture and CTPs are known producers of HA, so this co-expression cannot speak to the native CTP state in vivo. FACS sorting can be employed on fresh bone marrow aspirates labeled for CD146 and HA. Four fractions can be obtained after FACS sorting (CD146-/HA-, CD146-/HA+, CD146+/HA-, and CD146+/HA+) and cultured for CTPs. In addition, the purity of FACS-sorted cells is often higher than MS processing (purity in the HA+++ fraction was 54.3% using the

EasySepTM system). If colonies are found in the CD146+/HA+ fractions, this would support the hypothesis that native CTPs co-express HA and CD146.

In a bone defect greater than 1-2 mm in diameter, cells experience profoundly hypoxic conditions. The response of HA+++ CTPs to hypoxic conditions is unknown. In order to predict the survival and proliferative response of these cells to hypoxia, cells obtained after HA-based MS processing were cultured in normoxic conditions for bone (3% O2), hypoxic conditions (0.1% O2) and hyperoxic conditions (21% O2, the standard cell culture condition). Preliminary data indicates that the enrichment of CTPs is maintained over the BCM marrow when cultured at 0.1%, 3% and 21% O2. However, the BCM and

MS-processed fractions have a lower prevalence when cultured at 0.1% O2, indicating that hypoxic conditions negatively influence the colony forming efficiency of these cells, consistent with prior studies.7 This data was obtained with only 3 patients, and power analysis indicates that at least 6 patients need to be tested to obtain significance in

266 accordance with the historical patient variability and previous enrichment levels (3.9-fold

± 2.53) using MS-processing (with a power level of 0.8). Additional patients should be added to this cohort to fully examine the response of HA+++ CTPs to hypoxic conditions.

Not all of the cells isolated in the HA+++ fraction are CTPs, and other cells with retained

HA are separated with the CTPs. The identity and function of these cells is unknown.

Could these cells function as “helper” cells that encourage increased proliferation of

HA+++ CTPs? Do they relay signals to CTPs to readily express alkaline phosphatase? Or is the increased proliferation and osteoblastic differentiation seen in colonies derived from HA+++ CTPs reflective of the intrinsic properties of the subset of CTPs that are isolated? Future work to uncover the identity of the HA+++ cells will illuminate the potential for tissue repair of the HA+++ population as a whole. Endothelial progenitors, also found in low numbers in the bone marrow, could associate with HA and may be a very desirable population for implantation with CTPs. An increased rate of development of new vasculature in a defect site will enable the delivery of oxygen and nutrients to implanted cells and CTPs, boosting survival of these cells and enhancing the ability of these osteogenic cells proliferate, differentiate, and repair the defect site.

8.3 Summary:

8.3.1. The application of HA-positive CTPs for improved bone regeneration

Selection of cells based on HA using magnetic separation has been shown to enrich for

CTPs in the HA+++ fraction (Chapter 3). Addition of this enriched cell population to a bone graft (Chapter 7) was expected to increase the magnitude of new bone formation in

267 the canine femoral multi-defect model (CFMD) by two complementary mechanisms: first, an increased concentration of CTPs in the graft, and second, the elimination of the more numerous, non-osteogenic cells that do not contribute to new bone formation. These abundant, non-osteogenic cells may hamper new bone growth by competing with CTPs for the limited oxygen and nutrients available at the graft site. Decreasing the number of these competing cells may allow CTPs to remain viable deeper within the graft volume.

+ While there was a trend for greater bone volume in the MS/HA W2 grafts, this was not significant using microCT analysis. However, the percent of new bone formed in the

+ MS/HA W2 defect sites was significantly higher as assessed by quantitative histomorphometry. Histological analysis also indicated that the percent of the defect

+ occupied by vascular sinus area was significantly greater in the MS/HA W2 grafted sites, indicating a more advanced state of revascularization in these defect sites.

8.4 Future work:

8.4.1 Implementation of protocol changes

Both hBMA and MS-processed cells transplanted into the canine femoral multidefect model formed robust new bone and showed an advanced state of remodeling after 4 weeks. The advantage of the CFMD model is its ability to implant 4 scaffolds in uniform defects in a single animal, thus eliminating the effect of animal to animal variability, a complicating factor in many animal models. This also allows the minimization of the number of animals used for experimentation. However, the sub-critical defect size may not be challenging enough to detect differences within cell processing techniques. It is

268 hypothesized that advancing to a more challenging defect model will enable clearer assessment of the additive effects of marrow processing.

In order to move into to the clinically relevant segmental defect model in the goat, the

HA-based magnetic separation protocol will have to be validated using caprine marrow.

Hyaluronan is expected to be conserved throughout species (and indeed this was demonstrated with canine marrow), so it is expected that the HABP and magnetic labeling will maintain effectiveness.

The scale up protocol for the HMS should be updated to reflect the protocol development results in Chapter 5 before implementation in the caprine model. Preincubation of the

HABP, TAC, and magnetic nanobeads will shorten the required labeling time by 35 minutes. It will also reduce the primary antibody cost while maintaining CTP enrichment.

In addition, in Chapter 7 it was noted that the second wash step resulted in the loss of

- + CTPs in the HA W2 fraction from the HA W2 fraction. This wash can be eliminated, saving

10 minutes of retention time in the magnet and preserving an additional 6.6% of the CTP population.

8.4.2 The Chronic Segmental Defect (CSD) Caprine model:

The Muschler lab, in collaboration with the University of Minnesota and the Institute for

Surgical Research, has developed a chronic segmental bone defect model in the goat tibia that incorporates the loss of regional soft tissues and periosteum, local tissue scarring, and fibrous tissue induced by the use of a local defect spacer. It is intended to better

269 represent the complexity and severity of the actual biological environment in which clinical bone grafts currently fail.

Each animal undergoes two surgical procedures, a “Pre-procedure” in which a 5 cm tibial defect is created and preserved with the use of a polymethylmethacrylate (PMMA) spacer, followed by the “Treatment Procedure” in which various clinically relevant bone grafting treatment scenarios can be implemented.

During the preprocedure, the 5 cm defect is created and the tibia stabilized with an intramedullary interlocking nail. The newly created defect is filled with a PMMA spacer to induce a membrane to form at the defect site. At the treatment procedure, the PMMA spacer is removed and the defect is filled with goat mineralized cancellous allograft enhanced with bone marrow cells obtained from bone marrow excavation or enriched by the MS processing based on HA.

X-rays of the treated tibia will be taken immediately after each surgery while the goats are still under anesthesia, and then be repeated with the goats sedated 4, 8 and 12 weeks after the second surgery when the defects were treated. Sixteen weeks after the pretreatment surgery and 12 weeks after the treatment surgery, the goats will be humanely euthanized. The tibia will be harvested for further assessment of defect healing using microCT and histological analysis.

Six goats will be treated with MS-processed cells isolated based on HA expression and

270 six goats will be treated with excavated marrow. This experiment tests additive effect of

MS-processing to increase the number of CTPs and decrease the number of competing non-osteogenic cells in this more challenging defect compared to marrow alone.

8.4.3 Development of alternative methods for selection of CTPs based on HA:

Magnetic separation is only one strategy for selection of cells with cell-associated HA.

Development of a scaffold that presents a hylauronan binding protein (HABP) can be paired with selective retention methodology to enhance CTP retention in a scaffold used for bone regeneration. Development of a method for tethering of HABP on the surface of osteoconductive scaffolds provides selection of HA-positive cells directly into the scaffold that will be implanted into the defect site. In addition, selective retention methods (Chapter 6) can be used to add another level of selection for CTPs when loading marrow onto the scaffold. Selective retention enables the preferential binding of CTPs on the scaffold due to their property of rapid attachment to surfaces. Tuning the surface area available for HA-positive cell retention through HABP presentation with the osteoconductive surface area available for cell and CTP retention can produce grafts with different CTP and cell concentrations. Since the ideal number of CTPs and cells in a given graft site is unknown, different versions of the graft can be tested with respect to the concentration and prevalence of CTPs in the canine femoral defect model or other suitable biologically relevant animal models.

8.5 References

271 1. Midura R, Su S, Morcuende J et al. Parathyroid hormone rapidly stimulates hyaluronan synthesis by periosteal osteoblasts in the tibial diaphysis of the growing rat. J

Biol Chem 2003. 278:51462-51468

2. Avidgor A, Goichberg P, Shivtiel S et al. CD44 and hyaluronic acid cooperate with

SDF-1 in the trafficking of human stem/progenitor cells to bone marrow. Blood 2004.

103(8):2981-2989

3. Sundstrom G, Lofvenberg E, Hassan I et al. Localisation and distribution of hyaluronan in normal bone marrow matrix: a novel method to evaluate impending fibrosis? Eur J Haematol 2002. 68:194-202

4. Sundstrom G, Dahl I, Hultdin M et al. Bone marrow hyaluronan distribution in patients with acute myeloid leukemia Medical Oncology 2005. 22:71-78

5. Sacchetti B, Funari A, Michienzi S et al. Self-renewing osteoprogenitors in bone marrow sinusoids can organize a hematopoietic microenvironment. Cell 2007.

131(2):324-336

6. Shi S and Gronthos S. Perivascular niche of postnatal mesenchymal stem cells in human bone marrow and dental pulp. J Bone and Min Res 2003 18(4): 696-704

7. Villarruel S, Boehm C, Pennington M et al 2008 The effect of oxygen tension on the in vitro assay of human osteoblastic connective tissue progenitor cells J Orthop Res 26(10):

1390-1397

272